{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.22399944000139999,
  "eval_steps": 500,
  "global_step": 2400,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015101841517857095,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4088.0,
      "completions/mean_length": 606.2172241210938,
      "completions/mean_terminated_length": 552.70703125,
      "completions/min_length": 2.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 9.333310000058333e-05,
      "grad_norm": 0.1384008526802063,
      "learning_rate": 2e-07,
      "loss": -0.0147,
      "num_tokens": 40532698.0,
      "reward": 0.4949602484703064,
      "reward_std": 0.2629406154155731,
      "rewards/simpleverify_reward/mean": 0.4949602484703064,
      "rewards/simpleverify_reward/std": 0.4999788999557495,
      "step": 1
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00018666620000116666,
      "grad_norm": 0.13450804352760315,
      "learning_rate": 2e-07,
      "loss": 0.0449,
      "step": 2
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00027999930000175,
      "grad_norm": 0.1348700374364853,
      "learning_rate": 2e-07,
      "loss": 0.0278,
      "step": 3
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0003733324000023333,
      "grad_norm": 0.1597312092781067,
      "learning_rate": 2e-07,
      "loss": -0.0179,
      "step": 4
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0004666655000029167,
      "grad_norm": 0.1616683453321457,
      "learning_rate": 2e-07,
      "loss": 0.0235,
      "step": 5
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0005599986000035,
      "grad_norm": 0.1294235736131668,
      "learning_rate": 2e-07,
      "loss": -0.0039,
      "step": 6
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0006533317000040833,
      "grad_norm": 0.12849809229373932,
      "learning_rate": 2e-07,
      "loss": 0.0026,
      "step": 7
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0007466648000046666,
      "grad_norm": 0.12978479266166687,
      "learning_rate": 2e-07,
      "loss": -0.0078,
      "step": 8
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00083999790000525,
      "grad_norm": 0.13568826019763947,
      "learning_rate": 2e-07,
      "loss": 0.04,
      "step": 9
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0009333310000058334,
      "grad_norm": 0.14199160039424896,
      "learning_rate": 2e-07,
      "loss": 0.0345,
      "step": 10
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0010266641000064166,
      "grad_norm": 0.14209267497062683,
      "learning_rate": 2e-07,
      "loss": 0.0078,
      "step": 11
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.001119997200007,
      "grad_norm": 0.14880694448947906,
      "learning_rate": 2e-07,
      "loss": 0.0261,
      "step": 12
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0012133303000075833,
      "grad_norm": 0.14299704134464264,
      "learning_rate": 2e-07,
      "loss": -0.0346,
      "step": 13
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0013066634000081666,
      "grad_norm": 0.14938555657863617,
      "learning_rate": 2e-07,
      "loss": 0.0212,
      "step": 14
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00139999650000875,
      "grad_norm": 0.14810962975025177,
      "learning_rate": 2e-07,
      "loss": -0.0232,
      "step": 15
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0014933296000093333,
      "grad_norm": 0.14529141783714294,
      "learning_rate": 2e-07,
      "loss": 0.0658,
      "step": 16
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0015866627000099165,
      "grad_norm": 0.1455274224281311,
      "learning_rate": 2e-07,
      "loss": -0.0176,
      "step": 17
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0016799958000105,
      "grad_norm": 0.16608372330665588,
      "learning_rate": 2e-07,
      "loss": 0.0133,
      "step": 18
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0017733289000110833,
      "grad_norm": 0.15505066514015198,
      "learning_rate": 2e-07,
      "loss": 0.0161,
      "step": 19
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0018666620000116667,
      "grad_norm": 0.14904393255710602,
      "learning_rate": 2e-07,
      "loss": 0.0315,
      "step": 20
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00195999510001225,
      "grad_norm": 0.14434808492660522,
      "learning_rate": 2e-07,
      "loss": 0.035,
      "step": 21
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.002053328200012833,
      "grad_norm": 0.1571751832962036,
      "learning_rate": 2e-07,
      "loss": -0.017,
      "step": 22
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0021466613000134167,
      "grad_norm": 0.16357989609241486,
      "learning_rate": 2e-07,
      "loss": 0.0199,
      "step": 23
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.002239994400014,
      "grad_norm": 0.18290925025939941,
      "learning_rate": 2e-07,
      "loss": 0.0082,
      "step": 24
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.002333327500014583,
      "grad_norm": 0.1514730155467987,
      "learning_rate": 2e-07,
      "loss": 0.0354,
      "step": 25
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0024266606000151666,
      "grad_norm": 0.14194783568382263,
      "learning_rate": 2e-07,
      "loss": 0.0168,
      "step": 26
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00251999370001575,
      "grad_norm": 0.1521742343902588,
      "learning_rate": 2e-07,
      "loss": 0.0935,
      "step": 27
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.002613326800016333,
      "grad_norm": 0.13186417520046234,
      "learning_rate": 2e-07,
      "loss": -0.0332,
      "step": 28
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0027066599000169166,
      "grad_norm": 0.20116032660007477,
      "learning_rate": 2e-07,
      "loss": 0.006,
      "step": 29
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0027999930000175,
      "grad_norm": 0.1515357494354248,
      "learning_rate": 2e-07,
      "loss": -0.0256,
      "step": 30
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.002893326100018083,
      "grad_norm": 0.15172889828681946,
      "learning_rate": 2e-07,
      "loss": 0.019,
      "step": 31
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0029866592000186666,
      "grad_norm": 0.15013325214385986,
      "learning_rate": 2e-07,
      "loss": -0.0013,
      "step": 32
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00307999230001925,
      "grad_norm": 0.15168893337249756,
      "learning_rate": 2e-07,
      "loss": 0.0498,
      "step": 33
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.003173325400019833,
      "grad_norm": 0.18933169543743134,
      "learning_rate": 2e-07,
      "loss": 0.0524,
      "step": 34
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0032666585000204165,
      "grad_norm": 0.16392450034618378,
      "learning_rate": 2e-07,
      "loss": 0.0415,
      "step": 35
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.003359991600021,
      "grad_norm": 0.19064961373806,
      "learning_rate": 2e-07,
      "loss": 0.0346,
      "step": 36
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0034533247000215835,
      "grad_norm": 0.1580137312412262,
      "learning_rate": 2e-07,
      "loss": 0.0544,
      "step": 37
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0035466578000221665,
      "grad_norm": 0.19043877720832825,
      "learning_rate": 2e-07,
      "loss": 0.0709,
      "step": 38
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00363999090002275,
      "grad_norm": 0.204293355345726,
      "learning_rate": 2e-07,
      "loss": 0.0375,
      "step": 39
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0037333240000233334,
      "grad_norm": 0.16858668625354767,
      "learning_rate": 2e-07,
      "loss": -0.0164,
      "step": 40
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0038266571000239165,
      "grad_norm": 0.16858167946338654,
      "learning_rate": 2e-07,
      "loss": 0.0138,
      "step": 41
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0039199902000245,
      "grad_norm": 0.1629207730293274,
      "learning_rate": 2e-07,
      "loss": 0.0197,
      "step": 42
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.004013323300025083,
      "grad_norm": 0.21567024290561676,
      "learning_rate": 2e-07,
      "loss": 0.048,
      "step": 43
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.004106656400025666,
      "grad_norm": 0.22543172538280487,
      "learning_rate": 2e-07,
      "loss": 0.0087,
      "step": 44
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00419998950002625,
      "grad_norm": 0.1762998253107071,
      "learning_rate": 2e-07,
      "loss": 0.0206,
      "step": 45
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.004293322600026833,
      "grad_norm": 0.20864303410053253,
      "learning_rate": 2e-07,
      "loss": 0.0375,
      "step": 46
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.004386655700027417,
      "grad_norm": 0.19752885401248932,
      "learning_rate": 2e-07,
      "loss": 0.0183,
      "step": 47
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.004479988800028,
      "grad_norm": 0.1848761886358261,
      "learning_rate": 2e-07,
      "loss": 0.0478,
      "step": 48
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.004573321900028583,
      "grad_norm": 0.27787908911705017,
      "learning_rate": 2e-07,
      "loss": 0.0309,
      "step": 49
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.004666655000029166,
      "grad_norm": 0.2812020778656006,
      "learning_rate": 2e-07,
      "loss": 0.0198,
      "step": 50
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00475998810002975,
      "grad_norm": 0.21986417472362518,
      "learning_rate": 2e-07,
      "loss": 0.0374,
      "step": 51
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.004853321200030333,
      "grad_norm": 0.2621350884437561,
      "learning_rate": 2e-07,
      "loss": 0.021,
      "step": 52
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.004946654300030917,
      "grad_norm": 0.20501263439655304,
      "learning_rate": 2e-07,
      "loss": 0.0191,
      "step": 53
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0050399874000315,
      "grad_norm": 0.1719241738319397,
      "learning_rate": 2e-07,
      "loss": -0.0386,
      "step": 54
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.005133320500032083,
      "grad_norm": 0.2824048697948456,
      "learning_rate": 2e-07,
      "loss": -0.034,
      "step": 55
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.005226653600032666,
      "grad_norm": 0.2427029013633728,
      "learning_rate": 2e-07,
      "loss": -0.0367,
      "step": 56
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00531998670003325,
      "grad_norm": 0.22163940966129303,
      "learning_rate": 2e-07,
      "loss": 0.0107,
      "step": 57
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.005413319800033833,
      "grad_norm": 0.35862430930137634,
      "learning_rate": 2e-07,
      "loss": 0.0754,
      "step": 58
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.005506652900034417,
      "grad_norm": 0.2770756483078003,
      "learning_rate": 2e-07,
      "loss": 0.02,
      "step": 59
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.005599986000035,
      "grad_norm": 0.25597915053367615,
      "learning_rate": 2e-07,
      "loss": 0.0175,
      "step": 60
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.005693319100035584,
      "grad_norm": 0.2557786703109741,
      "learning_rate": 2e-07,
      "loss": -0.0224,
      "step": 61
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.005786652200036166,
      "grad_norm": 0.30547863245010376,
      "learning_rate": 2e-07,
      "loss": 0.0494,
      "step": 62
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00587998530003675,
      "grad_norm": 0.3096722364425659,
      "learning_rate": 2e-07,
      "loss": -0.0235,
      "step": 63
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.005973318400037333,
      "grad_norm": 0.26014670729637146,
      "learning_rate": 2e-07,
      "loss": 0.0719,
      "step": 64
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01416015625,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4083.0,
      "completions/mean_length": 602.6310424804688,
      "completions/mean_terminated_length": 552.453857421875,
      "completions/min_length": 4.0,
      "completions/min_terminated_length": 4.0,
      "epoch": 0.006066651500037917,
      "grad_norm": 0.1104929968714714,
      "learning_rate": 2e-07,
      "loss": 0.0383,
      "num_tokens": 80742274.0,
      "reward": 0.5538853406906128,
      "reward_std": 0.21083912253379822,
      "rewards/simpleverify_reward/mean": 0.5538853406906128,
      "rewards/simpleverify_reward/std": 0.49709224700927734,
      "step": 65
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0061599846000385,
      "grad_norm": 0.11229357123374939,
      "learning_rate": 2e-07,
      "loss": 0.0065,
      "step": 66
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0062533177000390835,
      "grad_norm": 0.12859536707401276,
      "learning_rate": 2e-07,
      "loss": 0.0236,
      "step": 67
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.006346650800039666,
      "grad_norm": 0.11728814244270325,
      "learning_rate": 2e-07,
      "loss": 0.0706,
      "step": 68
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00643998390004025,
      "grad_norm": 0.13092705607414246,
      "learning_rate": 2e-07,
      "loss": 0.0183,
      "step": 69
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.006533317000040833,
      "grad_norm": 0.11778075993061066,
      "learning_rate": 2e-07,
      "loss": 0.0162,
      "step": 70
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0066266501000414165,
      "grad_norm": 0.1319500207901001,
      "learning_rate": 2e-07,
      "loss": 0.0406,
      "step": 71
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.006719983200042,
      "grad_norm": 0.11401458084583282,
      "learning_rate": 2e-07,
      "loss": -0.0647,
      "step": 72
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0068133163000425835,
      "grad_norm": 0.11675283312797546,
      "learning_rate": 2e-07,
      "loss": 0.0509,
      "step": 73
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.006906649400043167,
      "grad_norm": 0.12296499311923981,
      "learning_rate": 2e-07,
      "loss": 0.0684,
      "step": 74
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0069999825000437495,
      "grad_norm": 0.1141868457198143,
      "learning_rate": 2e-07,
      "loss": 0.0491,
      "step": 75
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.007093315600044333,
      "grad_norm": 0.11803154647350311,
      "learning_rate": 2e-07,
      "loss": -0.0113,
      "step": 76
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0071866487000449165,
      "grad_norm": 0.12159748375415802,
      "learning_rate": 2e-07,
      "loss": 0.0246,
      "step": 77
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0072799818000455,
      "grad_norm": 0.13280890882015228,
      "learning_rate": 2e-07,
      "loss": 0.044,
      "step": 78
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.007373314900046083,
      "grad_norm": 0.11563581228256226,
      "learning_rate": 2e-07,
      "loss": 0.0591,
      "step": 79
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.007466648000046667,
      "grad_norm": 0.12397162616252899,
      "learning_rate": 2e-07,
      "loss": 0.0293,
      "step": 80
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0075599811000472495,
      "grad_norm": 0.12922480702400208,
      "learning_rate": 2e-07,
      "loss": 0.0283,
      "step": 81
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.007653314200047833,
      "grad_norm": 0.11023665219545364,
      "learning_rate": 2e-07,
      "loss": 0.0284,
      "step": 82
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.007746647300048416,
      "grad_norm": 0.11985830962657928,
      "learning_rate": 2e-07,
      "loss": 0.0829,
      "step": 83
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.007839980400049,
      "grad_norm": 0.1204838901758194,
      "learning_rate": 2e-07,
      "loss": 0.0369,
      "step": 84
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.007933313500049582,
      "grad_norm": 0.12884598970413208,
      "learning_rate": 2e-07,
      "loss": -0.0038,
      "step": 85
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.008026646600050166,
      "grad_norm": 0.11503426730632782,
      "learning_rate": 2e-07,
      "loss": 0.0121,
      "step": 86
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00811997970005075,
      "grad_norm": 0.1093788594007492,
      "learning_rate": 2e-07,
      "loss": 0.046,
      "step": 87
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.008213312800051333,
      "grad_norm": 0.1122165322303772,
      "learning_rate": 2e-07,
      "loss": 0.0541,
      "step": 88
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.008306645900051916,
      "grad_norm": 0.11874477565288544,
      "learning_rate": 2e-07,
      "loss": -0.024,
      "step": 89
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0083999790000525,
      "grad_norm": 0.12048187106847763,
      "learning_rate": 2e-07,
      "loss": 0.0459,
      "step": 90
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.008493312100053083,
      "grad_norm": 0.13318213820457458,
      "learning_rate": 2e-07,
      "loss": 0.0106,
      "step": 91
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.008586645200053667,
      "grad_norm": 0.13598953187465668,
      "learning_rate": 2e-07,
      "loss": 0.0471,
      "step": 92
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00867997830005425,
      "grad_norm": 0.1182006448507309,
      "learning_rate": 2e-07,
      "loss": 0.0114,
      "step": 93
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.008773311400054834,
      "grad_norm": 0.1281680166721344,
      "learning_rate": 2e-07,
      "loss": 0.0171,
      "step": 94
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.008866644500055417,
      "grad_norm": 0.12943843007087708,
      "learning_rate": 2e-07,
      "loss": -0.0092,
      "step": 95
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.008959977600056,
      "grad_norm": 0.11512872576713562,
      "learning_rate": 2e-07,
      "loss": 0.0429,
      "step": 96
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.009053310700056582,
      "grad_norm": 0.12001234292984009,
      "learning_rate": 2e-07,
      "loss": -0.0073,
      "step": 97
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.009146643800057166,
      "grad_norm": 0.1210739016532898,
      "learning_rate": 2e-07,
      "loss": 0.0295,
      "step": 98
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00923997690005775,
      "grad_norm": 0.124770887196064,
      "learning_rate": 2e-07,
      "loss": 0.02,
      "step": 99
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.009333310000058333,
      "grad_norm": 0.11245358735322952,
      "learning_rate": 2e-07,
      "loss": 0.0193,
      "step": 100
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.009426643100058916,
      "grad_norm": 0.11386345326900482,
      "learning_rate": 2e-07,
      "loss": 0.0639,
      "step": 101
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0095199762000595,
      "grad_norm": 0.12418945878744125,
      "learning_rate": 2e-07,
      "loss": 0.0139,
      "step": 102
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.009613309300060083,
      "grad_norm": 0.12058115750551224,
      "learning_rate": 2e-07,
      "loss": 0.0484,
      "step": 103
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.009706642400060667,
      "grad_norm": 0.12680678069591522,
      "learning_rate": 2e-07,
      "loss": 0.002,
      "step": 104
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.00979997550006125,
      "grad_norm": 0.13684795796871185,
      "learning_rate": 2e-07,
      "loss": 0.0507,
      "step": 105
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.009893308600061834,
      "grad_norm": 0.11940007656812668,
      "learning_rate": 2e-07,
      "loss": 0.0325,
      "step": 106
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.009986641700062417,
      "grad_norm": 0.11589401960372925,
      "learning_rate": 2e-07,
      "loss": 0.0425,
      "step": 107
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.010079974800063,
      "grad_norm": 0.129934161901474,
      "learning_rate": 2e-07,
      "loss": 0.0255,
      "step": 108
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.010173307900063584,
      "grad_norm": 0.1248784214258194,
      "learning_rate": 2e-07,
      "loss": 0.0002,
      "step": 109
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.010266641000064166,
      "grad_norm": 0.12240680307149887,
      "learning_rate": 2e-07,
      "loss": 0.0264,
      "step": 110
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.010359974100064749,
      "grad_norm": 0.12803420424461365,
      "learning_rate": 2e-07,
      "loss": 0.0163,
      "step": 111
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.010453307200065333,
      "grad_norm": 0.11498578637838364,
      "learning_rate": 2e-07,
      "loss": 0.0484,
      "step": 112
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.010546640300065916,
      "grad_norm": 0.11552631109952927,
      "learning_rate": 2e-07,
      "loss": 0.0551,
      "step": 113
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0106399734000665,
      "grad_norm": 0.11569365859031677,
      "learning_rate": 2e-07,
      "loss": 0.0066,
      "step": 114
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.010733306500067083,
      "grad_norm": 0.11575288325548172,
      "learning_rate": 2e-07,
      "loss": 0.0326,
      "step": 115
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.010826639600067666,
      "grad_norm": 0.11555801331996918,
      "learning_rate": 2e-07,
      "loss": -0.0182,
      "step": 116
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01091997270006825,
      "grad_norm": 0.12371595948934555,
      "learning_rate": 2e-07,
      "loss": -0.0353,
      "step": 117
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.011013305800068833,
      "grad_norm": 0.11161545664072037,
      "learning_rate": 2e-07,
      "loss": -0.0043,
      "step": 118
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.011106638900069417,
      "grad_norm": 0.125113844871521,
      "learning_rate": 2e-07,
      "loss": 0.0181,
      "step": 119
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01119997200007,
      "grad_norm": 0.12157007306814194,
      "learning_rate": 2e-07,
      "loss": -0.01,
      "step": 120
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.011293305100070584,
      "grad_norm": 0.12889772653579712,
      "learning_rate": 2e-07,
      "loss": 0.0606,
      "step": 121
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.011386638200071167,
      "grad_norm": 0.12442421913146973,
      "learning_rate": 2e-07,
      "loss": 0.0378,
      "step": 122
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.011479971300071749,
      "grad_norm": 0.12123935669660568,
      "learning_rate": 2e-07,
      "loss": 0.0293,
      "step": 123
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.011573304400072332,
      "grad_norm": 0.11454013735055923,
      "learning_rate": 2e-07,
      "loss": -0.033,
      "step": 124
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.011666637500072916,
      "grad_norm": 0.14208470284938812,
      "learning_rate": 2e-07,
      "loss": 0.0569,
      "step": 125
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0117599706000735,
      "grad_norm": 0.13243162631988525,
      "learning_rate": 2e-07,
      "loss": -0.0346,
      "step": 126
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.011853303700074083,
      "grad_norm": 0.11157102137804031,
      "learning_rate": 2e-07,
      "loss": 0.0,
      "step": 127
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.011946636800074666,
      "grad_norm": 0.13012441992759705,
      "learning_rate": 2e-07,
      "loss": 0.011,
      "step": 128
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.013863699776785698,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4090.0,
      "completions/mean_length": 605.7926025390625,
      "completions/mean_terminated_length": 556.72509765625,
      "completions/min_length": 14.0,
      "completions/min_terminated_length": 14.0,
      "epoch": 0.01203996990007525,
      "grad_norm": 0.12900763750076294,
      "learning_rate": 2e-07,
      "loss": 0.0005,
      "num_tokens": 121267841.0,
      "reward": 0.552001953125,
      "reward_std": 0.21141129732131958,
      "rewards/simpleverify_reward/mean": 0.552001953125,
      "rewards/simpleverify_reward/std": 0.49729272723197937,
      "step": 129
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.012133303000075833,
      "grad_norm": 0.11976109445095062,
      "learning_rate": 2e-07,
      "loss": 0.0134,
      "step": 130
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.012226636100076417,
      "grad_norm": 0.1242467537522316,
      "learning_rate": 2e-07,
      "loss": -0.0174,
      "step": 131
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.012319969200077,
      "grad_norm": 0.1271137148141861,
      "learning_rate": 2e-07,
      "loss": 0.0942,
      "step": 132
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.012413302300077584,
      "grad_norm": 0.12402600049972534,
      "learning_rate": 2e-07,
      "loss": -0.0061,
      "step": 133
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.012506635400078167,
      "grad_norm": 0.115492083132267,
      "learning_rate": 2e-07,
      "loss": 0.0171,
      "step": 134
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01259996850007875,
      "grad_norm": 0.12605193257331848,
      "learning_rate": 2e-07,
      "loss": 0.0221,
      "step": 135
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.012693301600079332,
      "grad_norm": 0.10870201885700226,
      "learning_rate": 2e-07,
      "loss": 0.0135,
      "step": 136
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.012786634700079916,
      "grad_norm": 0.11895795166492462,
      "learning_rate": 2e-07,
      "loss": 0.0295,
      "step": 137
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0128799678000805,
      "grad_norm": 0.1160581037402153,
      "learning_rate": 2e-07,
      "loss": -0.0232,
      "step": 138
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.012973300900081083,
      "grad_norm": 0.12504318356513977,
      "learning_rate": 2e-07,
      "loss": 0.0529,
      "step": 139
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.013066634000081666,
      "grad_norm": 0.11358392238616943,
      "learning_rate": 2e-07,
      "loss": 0.0407,
      "step": 140
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01315996710008225,
      "grad_norm": 0.11238788813352585,
      "learning_rate": 2e-07,
      "loss": 0.0048,
      "step": 141
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.013253300200082833,
      "grad_norm": 0.11288043856620789,
      "learning_rate": 2e-07,
      "loss": 0.012,
      "step": 142
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.013346633300083417,
      "grad_norm": 0.12846171855926514,
      "learning_rate": 2e-07,
      "loss": -0.0201,
      "step": 143
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.013439966400084,
      "grad_norm": 0.1235656589269638,
      "learning_rate": 2e-07,
      "loss": 0.0596,
      "step": 144
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.013533299500084583,
      "grad_norm": 0.1266559213399887,
      "learning_rate": 2e-07,
      "loss": 0.0439,
      "step": 145
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.013626632600085167,
      "grad_norm": 0.11292146891355515,
      "learning_rate": 2e-07,
      "loss": 0.0322,
      "step": 146
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01371996570008575,
      "grad_norm": 0.12272363156080246,
      "learning_rate": 2e-07,
      "loss": 0.0592,
      "step": 147
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.013813298800086334,
      "grad_norm": 0.11472923308610916,
      "learning_rate": 2e-07,
      "loss": 0.0068,
      "step": 148
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.013906631900086916,
      "grad_norm": 0.11798423528671265,
      "learning_rate": 2e-07,
      "loss": 0.0426,
      "step": 149
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.013999965000087499,
      "grad_norm": 0.12502296268939972,
      "learning_rate": 2e-07,
      "loss": 0.0166,
      "step": 150
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.014093298100088083,
      "grad_norm": 0.11589020490646362,
      "learning_rate": 2e-07,
      "loss": 0.0337,
      "step": 151
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.014186631200088666,
      "grad_norm": 0.12701231241226196,
      "learning_rate": 2e-07,
      "loss": 0.0482,
      "step": 152
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01427996430008925,
      "grad_norm": 0.10634884983301163,
      "learning_rate": 2e-07,
      "loss": 0.0042,
      "step": 153
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.014373297400089833,
      "grad_norm": 0.11809704452753067,
      "learning_rate": 2e-07,
      "loss": 0.0196,
      "step": 154
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.014466630500090416,
      "grad_norm": 0.13478153944015503,
      "learning_rate": 2e-07,
      "loss": 0.0058,
      "step": 155
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.014559963600091,
      "grad_norm": 0.1195162758231163,
      "learning_rate": 2e-07,
      "loss": 0.056,
      "step": 156
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.014653296700091583,
      "grad_norm": 0.11595071107149124,
      "learning_rate": 2e-07,
      "loss": 0.0649,
      "step": 157
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.014746629800092167,
      "grad_norm": 0.12632755935192108,
      "learning_rate": 2e-07,
      "loss": -0.0015,
      "step": 158
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01483996290009275,
      "grad_norm": 0.11408816277980804,
      "learning_rate": 2e-07,
      "loss": 0.051,
      "step": 159
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.014933296000093334,
      "grad_norm": 0.1233525425195694,
      "learning_rate": 2e-07,
      "loss": 0.0814,
      "step": 160
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.015026629100093917,
      "grad_norm": 0.11964831501245499,
      "learning_rate": 2e-07,
      "loss": -0.0034,
      "step": 161
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.015119962200094499,
      "grad_norm": 0.11419013142585754,
      "learning_rate": 2e-07,
      "loss": 0.0566,
      "step": 162
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.015213295300095082,
      "grad_norm": 0.11575242877006531,
      "learning_rate": 2e-07,
      "loss": 0.0263,
      "step": 163
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.015306628400095666,
      "grad_norm": 0.11876792460680008,
      "learning_rate": 2e-07,
      "loss": -0.0166,
      "step": 164
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01539996150009625,
      "grad_norm": 0.12759201228618622,
      "learning_rate": 2e-07,
      "loss": 0.0211,
      "step": 165
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.015493294600096833,
      "grad_norm": 0.12791675329208374,
      "learning_rate": 2e-07,
      "loss": 0.0246,
      "step": 166
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.015586627700097416,
      "grad_norm": 0.11425651609897614,
      "learning_rate": 2e-07,
      "loss": -0.0075,
      "step": 167
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.015679960800098,
      "grad_norm": 0.12347857654094696,
      "learning_rate": 2e-07,
      "loss": 0.0169,
      "step": 168
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.015773293900098583,
      "grad_norm": 0.1158008798956871,
      "learning_rate": 2e-07,
      "loss": -0.0052,
      "step": 169
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.015866627000099165,
      "grad_norm": 0.12241604924201965,
      "learning_rate": 2e-07,
      "loss": 0.0353,
      "step": 170
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01595996010009975,
      "grad_norm": 0.11978351324796677,
      "learning_rate": 2e-07,
      "loss": 0.0497,
      "step": 171
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.016053293200100332,
      "grad_norm": 0.11186036467552185,
      "learning_rate": 2e-07,
      "loss": -0.0239,
      "step": 172
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.016146626300100917,
      "grad_norm": 0.11632370948791504,
      "learning_rate": 2e-07,
      "loss": 0.0577,
      "step": 173
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0162399594001015,
      "grad_norm": 0.12881873548030853,
      "learning_rate": 2e-07,
      "loss": 0.0534,
      "step": 174
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.016333292500102084,
      "grad_norm": 0.11721882969141006,
      "learning_rate": 2e-07,
      "loss": -0.018,
      "step": 175
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.016426625600102666,
      "grad_norm": 0.1351327896118164,
      "learning_rate": 2e-07,
      "loss": -0.0178,
      "step": 176
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01651995870010325,
      "grad_norm": 0.11248568445444107,
      "learning_rate": 2e-07,
      "loss": -0.0057,
      "step": 177
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.016613291800103833,
      "grad_norm": 0.11764824390411377,
      "learning_rate": 2e-07,
      "loss": 0.0408,
      "step": 178
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.016706624900104418,
      "grad_norm": 0.13204790651798248,
      "learning_rate": 2e-07,
      "loss": 0.0438,
      "step": 179
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.016799958000105,
      "grad_norm": 0.12130002677440643,
      "learning_rate": 2e-07,
      "loss": 0.0207,
      "step": 180
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.016893291100105585,
      "grad_norm": 0.11385934054851532,
      "learning_rate": 2e-07,
      "loss": 0.0587,
      "step": 181
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.016986624200106167,
      "grad_norm": 0.11729420721530914,
      "learning_rate": 2e-07,
      "loss": 0.0419,
      "step": 182
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.017079957300106748,
      "grad_norm": 0.11200563609600067,
      "learning_rate": 2e-07,
      "loss": 0.0386,
      "step": 183
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.017173290400107333,
      "grad_norm": 0.12005419284105301,
      "learning_rate": 2e-07,
      "loss": -0.0178,
      "step": 184
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.017266623500107915,
      "grad_norm": 0.1239481270313263,
      "learning_rate": 2e-07,
      "loss": 0.042,
      "step": 185
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0173599566001085,
      "grad_norm": 0.11413433402776718,
      "learning_rate": 2e-07,
      "loss": 0.0456,
      "step": 186
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.017453289700109082,
      "grad_norm": 0.12707901000976562,
      "learning_rate": 2e-07,
      "loss": 0.0366,
      "step": 187
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.017546622800109667,
      "grad_norm": 0.12394179403781891,
      "learning_rate": 2e-07,
      "loss": 0.0301,
      "step": 188
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01763995590011025,
      "grad_norm": 0.1205565333366394,
      "learning_rate": 2e-07,
      "loss": 0.0374,
      "step": 189
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.017733289000110834,
      "grad_norm": 0.11749666929244995,
      "learning_rate": 2e-07,
      "loss": -0.0035,
      "step": 190
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.017826622100111416,
      "grad_norm": 0.11596448719501495,
      "learning_rate": 2e-07,
      "loss": 0.0627,
      "step": 191
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.017919955200112,
      "grad_norm": 0.12997643649578094,
      "learning_rate": 2e-07,
      "loss": 0.0235,
      "step": 192
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.014439174107142905,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4089.0,
      "completions/mean_length": 607.7674560546875,
      "completions/mean_terminated_length": 556.6622924804688,
      "completions/min_length": 19.0,
      "completions/min_terminated_length": 19.0,
      "epoch": 0.018013288300112583,
      "grad_norm": 0.12148799747228622,
      "learning_rate": 2e-07,
      "loss": 0.0042,
      "num_tokens": 161796159.0,
      "reward": 0.5532052516937256,
      "reward_std": 0.20556531846523285,
      "rewards/simpleverify_reward/mean": 0.5532051920890808,
      "rewards/simpleverify_reward/std": 0.4971654415130615,
      "step": 193
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.018106621400113165,
      "grad_norm": 0.11413375288248062,
      "learning_rate": 2e-07,
      "loss": 0.0473,
      "step": 194
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01819995450011375,
      "grad_norm": 0.12304969131946564,
      "learning_rate": 2e-07,
      "loss": 0.0029,
      "step": 195
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01829328760011433,
      "grad_norm": 0.1279825121164322,
      "learning_rate": 2e-07,
      "loss": -0.0084,
      "step": 196
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.018386620700114917,
      "grad_norm": 0.10730310529470444,
      "learning_rate": 2e-07,
      "loss": -0.0079,
      "step": 197
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0184799538001155,
      "grad_norm": 0.12900662422180176,
      "learning_rate": 2e-07,
      "loss": 0.0582,
      "step": 198
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.018573286900116084,
      "grad_norm": 0.1261545717716217,
      "learning_rate": 2e-07,
      "loss": 0.0585,
      "step": 199
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.018666620000116665,
      "grad_norm": 0.1182534247636795,
      "learning_rate": 2e-07,
      "loss": 0.0575,
      "step": 200
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01875995310011725,
      "grad_norm": 0.11720357090234756,
      "learning_rate": 2e-07,
      "loss": 0.016,
      "step": 201
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.018853286200117832,
      "grad_norm": 0.12043602764606476,
      "learning_rate": 2e-07,
      "loss": 0.0563,
      "step": 202
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.018946619300118418,
      "grad_norm": 0.11089618504047394,
      "learning_rate": 2e-07,
      "loss": 0.0261,
      "step": 203
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.019039952400119,
      "grad_norm": 0.10863696038722992,
      "learning_rate": 2e-07,
      "loss": 0.0051,
      "step": 204
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.019133285500119585,
      "grad_norm": 0.11304053664207458,
      "learning_rate": 2e-07,
      "loss": 0.0508,
      "step": 205
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.019226618600120166,
      "grad_norm": 0.11691603809595108,
      "learning_rate": 2e-07,
      "loss": -0.0158,
      "step": 206
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.019319951700120748,
      "grad_norm": 0.1256258636713028,
      "learning_rate": 2e-07,
      "loss": 0.0878,
      "step": 207
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.019413284800121333,
      "grad_norm": 0.12094900012016296,
      "learning_rate": 2e-07,
      "loss": 0.0259,
      "step": 208
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.019506617900121915,
      "grad_norm": 0.13027654588222504,
      "learning_rate": 2e-07,
      "loss": 0.0189,
      "step": 209
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0195999510001225,
      "grad_norm": 0.1252318024635315,
      "learning_rate": 2e-07,
      "loss": 0.0193,
      "step": 210
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.019693284100123082,
      "grad_norm": 0.11989153176546097,
      "learning_rate": 2e-07,
      "loss": 0.0605,
      "step": 211
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.019786617200123667,
      "grad_norm": 0.1289915293455124,
      "learning_rate": 2e-07,
      "loss": -0.0137,
      "step": 212
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.01987995030012425,
      "grad_norm": 0.11391989886760712,
      "learning_rate": 2e-07,
      "loss": 0.0234,
      "step": 213
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.019973283400124834,
      "grad_norm": 0.11034820228815079,
      "learning_rate": 2e-07,
      "loss": 0.0079,
      "step": 214
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.020066616500125416,
      "grad_norm": 0.12173809111118317,
      "learning_rate": 2e-07,
      "loss": 0.007,
      "step": 215
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.020159949600126,
      "grad_norm": 0.11439619958400726,
      "learning_rate": 2e-07,
      "loss": -0.0044,
      "step": 216
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.020253282700126583,
      "grad_norm": 0.12216250598430634,
      "learning_rate": 2e-07,
      "loss": -0.0015,
      "step": 217
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.020346615800127168,
      "grad_norm": 0.11854705214500427,
      "learning_rate": 2e-07,
      "loss": -0.0137,
      "step": 218
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02043994890012775,
      "grad_norm": 0.11740019917488098,
      "learning_rate": 2e-07,
      "loss": 0.0347,
      "step": 219
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02053328200012833,
      "grad_norm": 0.11064712703227997,
      "learning_rate": 2e-07,
      "loss": 0.0802,
      "step": 220
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.020626615100128916,
      "grad_norm": 0.1068812757730484,
      "learning_rate": 2e-07,
      "loss": 0.0338,
      "step": 221
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.020719948200129498,
      "grad_norm": 0.11171789467334747,
      "learning_rate": 2e-07,
      "loss": -0.0278,
      "step": 222
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.020813281300130083,
      "grad_norm": 0.12349051237106323,
      "learning_rate": 2e-07,
      "loss": -0.0184,
      "step": 223
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.020906614400130665,
      "grad_norm": 0.1268196702003479,
      "learning_rate": 2e-07,
      "loss": -0.0025,
      "step": 224
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02099994750013125,
      "grad_norm": 0.11517728120088577,
      "learning_rate": 2e-07,
      "loss": 0.0095,
      "step": 225
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.021093280600131832,
      "grad_norm": 0.11174178868532181,
      "learning_rate": 2e-07,
      "loss": 0.0186,
      "step": 226
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.021186613700132417,
      "grad_norm": 0.12372954189777374,
      "learning_rate": 2e-07,
      "loss": 0.0155,
      "step": 227
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.021279946800133,
      "grad_norm": 0.1380620300769806,
      "learning_rate": 2e-07,
      "loss": -0.0025,
      "step": 228
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.021373279900133584,
      "grad_norm": 0.12020408362150192,
      "learning_rate": 2e-07,
      "loss": 0.0184,
      "step": 229
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.021466613000134166,
      "grad_norm": 0.10718788951635361,
      "learning_rate": 2e-07,
      "loss": 0.0039,
      "step": 230
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02155994610013475,
      "grad_norm": 0.11507224291563034,
      "learning_rate": 2e-07,
      "loss": 0.0373,
      "step": 231
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.021653279200135333,
      "grad_norm": 0.12471979856491089,
      "learning_rate": 2e-07,
      "loss": 0.0121,
      "step": 232
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.021746612300135915,
      "grad_norm": 0.1148390918970108,
      "learning_rate": 2e-07,
      "loss": 0.0884,
      "step": 233
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0218399454001365,
      "grad_norm": 0.10520583391189575,
      "learning_rate": 2e-07,
      "loss": -0.0148,
      "step": 234
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02193327850013708,
      "grad_norm": 0.12078914046287537,
      "learning_rate": 2e-07,
      "loss": 0.0656,
      "step": 235
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.022026611600137667,
      "grad_norm": 0.12394321709871292,
      "learning_rate": 2e-07,
      "loss": 0.0058,
      "step": 236
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02211994470013825,
      "grad_norm": 0.11081356555223465,
      "learning_rate": 2e-07,
      "loss": 0.0124,
      "step": 237
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.022213277800138834,
      "grad_norm": 0.13361530005931854,
      "learning_rate": 2e-07,
      "loss": 0.0723,
      "step": 238
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.022306610900139415,
      "grad_norm": 0.11601588129997253,
      "learning_rate": 2e-07,
      "loss": 0.0673,
      "step": 239
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02239994400014,
      "grad_norm": 0.1133677214384079,
      "learning_rate": 2e-07,
      "loss": 0.0653,
      "step": 240
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.022493277100140582,
      "grad_norm": 0.11076609790325165,
      "learning_rate": 2e-07,
      "loss": 0.0399,
      "step": 241
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.022586610200141168,
      "grad_norm": 0.11817794293165207,
      "learning_rate": 2e-07,
      "loss": 0.0336,
      "step": 242
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02267994330014175,
      "grad_norm": 0.11222653090953827,
      "learning_rate": 2e-07,
      "loss": 0.0529,
      "step": 243
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.022773276400142334,
      "grad_norm": 0.11679072678089142,
      "learning_rate": 2e-07,
      "loss": 0.0135,
      "step": 244
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.022866609500142916,
      "grad_norm": 0.11334266513586044,
      "learning_rate": 2e-07,
      "loss": 0.0316,
      "step": 245
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.022959942600143498,
      "grad_norm": 0.11308363825082779,
      "learning_rate": 2e-07,
      "loss": 0.0573,
      "step": 246
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.023053275700144083,
      "grad_norm": 0.12372971326112747,
      "learning_rate": 2e-07,
      "loss": 0.082,
      "step": 247
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.023146608800144665,
      "grad_norm": 0.12267614156007767,
      "learning_rate": 2e-07,
      "loss": -0.0151,
      "step": 248
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02323994190014525,
      "grad_norm": 0.11206039041280746,
      "learning_rate": 2e-07,
      "loss": 0.0176,
      "step": 249
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.023333275000145832,
      "grad_norm": 0.12499067932367325,
      "learning_rate": 2e-07,
      "loss": 0.0357,
      "step": 250
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.023426608100146417,
      "grad_norm": 0.12153958529233932,
      "learning_rate": 2e-07,
      "loss": 0.0173,
      "step": 251
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.023519941200147,
      "grad_norm": 0.12343767285346985,
      "learning_rate": 2e-07,
      "loss": 0.0651,
      "step": 252
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.023613274300147584,
      "grad_norm": 0.11580204218626022,
      "learning_rate": 2e-07,
      "loss": 0.0082,
      "step": 253
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.023706607400148166,
      "grad_norm": 0.15583446621894836,
      "learning_rate": 2e-07,
      "loss": 0.0155,
      "step": 254
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02379994050014875,
      "grad_norm": 0.11579901725053787,
      "learning_rate": 2e-07,
      "loss": 0.0042,
      "step": 255
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.023893273600149333,
      "grad_norm": 0.10898569226264954,
      "learning_rate": 2e-07,
      "loss": 0.0009,
      "step": 256
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.014212472098214302,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4089.0,
      "completions/mean_length": 606.5999145507812,
      "completions/mean_terminated_length": 556.2919311523438,
      "completions/min_length": 5.0,
      "completions/min_terminated_length": 5.0,
      "epoch": 0.023986606700149918,
      "grad_norm": 0.12543149292469025,
      "learning_rate": 2e-07,
      "loss": 0.0598,
      "num_tokens": 202181815.0,
      "reward": 0.5600237250328064,
      "reward_std": 0.20419713854789734,
      "rewards/simpleverify_reward/mean": 0.5600237250328064,
      "rewards/simpleverify_reward/std": 0.4963884651660919,
      "step": 257
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0240799398001505,
      "grad_norm": 0.11861643195152283,
      "learning_rate": 2e-07,
      "loss": 0.0575,
      "step": 258
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02417327290015108,
      "grad_norm": 0.11964251101016998,
      "learning_rate": 2e-07,
      "loss": 0.0071,
      "step": 259
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.024266606000151666,
      "grad_norm": 0.11687079817056656,
      "learning_rate": 2e-07,
      "loss": 0.0029,
      "step": 260
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.024359939100152248,
      "grad_norm": 0.12559251487255096,
      "learning_rate": 2e-07,
      "loss": 0.0233,
      "step": 261
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.024453272200152833,
      "grad_norm": 0.11664198338985443,
      "learning_rate": 2e-07,
      "loss": 0.0217,
      "step": 262
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.024546605300153415,
      "grad_norm": 0.12424785643815994,
      "learning_rate": 2e-07,
      "loss": 0.0445,
      "step": 263
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.024639938400154,
      "grad_norm": 0.11672412604093552,
      "learning_rate": 2e-07,
      "loss": 0.0064,
      "step": 264
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.024733271500154582,
      "grad_norm": 0.11804558336734772,
      "learning_rate": 2e-07,
      "loss": 0.0397,
      "step": 265
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.024826604600155167,
      "grad_norm": 0.11709222942590714,
      "learning_rate": 2e-07,
      "loss": 0.0379,
      "step": 266
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02491993770015575,
      "grad_norm": 0.12235447019338608,
      "learning_rate": 2e-07,
      "loss": -0.0187,
      "step": 267
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.025013270800156334,
      "grad_norm": 0.12662118673324585,
      "learning_rate": 2e-07,
      "loss": 0.0304,
      "step": 268
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.025106603900156916,
      "grad_norm": 0.10628720372915268,
      "learning_rate": 2e-07,
      "loss": 0.0409,
      "step": 269
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0251999370001575,
      "grad_norm": 0.13236229121685028,
      "learning_rate": 2e-07,
      "loss": -0.0009,
      "step": 270
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.025293270100158083,
      "grad_norm": 0.10969637334346771,
      "learning_rate": 2e-07,
      "loss": -0.0023,
      "step": 271
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.025386603200158665,
      "grad_norm": 0.11970086395740509,
      "learning_rate": 2e-07,
      "loss": 0.0604,
      "step": 272
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02547993630015925,
      "grad_norm": 0.1275329291820526,
      "learning_rate": 2e-07,
      "loss": 0.0549,
      "step": 273
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02557326940015983,
      "grad_norm": 0.136390820145607,
      "learning_rate": 2e-07,
      "loss": 0.0522,
      "step": 274
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.025666602500160417,
      "grad_norm": 0.11464012414216995,
      "learning_rate": 2e-07,
      "loss": 0.0298,
      "step": 275
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.025759935600161,
      "grad_norm": 0.12442323565483093,
      "learning_rate": 2e-07,
      "loss": 0.0481,
      "step": 276
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.025853268700161584,
      "grad_norm": 0.12034337222576141,
      "learning_rate": 2e-07,
      "loss": 0.0217,
      "step": 277
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.025946601800162165,
      "grad_norm": 0.12065837532281876,
      "learning_rate": 2e-07,
      "loss": 0.0426,
      "step": 278
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02603993490016275,
      "grad_norm": 0.1317777931690216,
      "learning_rate": 2e-07,
      "loss": -0.0177,
      "step": 279
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.026133268000163332,
      "grad_norm": 0.12822996079921722,
      "learning_rate": 2e-07,
      "loss": -0.0009,
      "step": 280
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.026226601100163918,
      "grad_norm": 0.11570777744054794,
      "learning_rate": 2e-07,
      "loss": 0.0552,
      "step": 281
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0263199342001645,
      "grad_norm": 0.12981736660003662,
      "learning_rate": 2e-07,
      "loss": 0.0368,
      "step": 282
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.026413267300165084,
      "grad_norm": 0.12223238497972488,
      "learning_rate": 2e-07,
      "loss": -0.0233,
      "step": 283
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.026506600400165666,
      "grad_norm": 0.11930467188358307,
      "learning_rate": 2e-07,
      "loss": 0.0127,
      "step": 284
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.026599933500166248,
      "grad_norm": 0.11253789812326431,
      "learning_rate": 2e-07,
      "loss": 0.0343,
      "step": 285
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.026693266600166833,
      "grad_norm": 0.10800273716449738,
      "learning_rate": 2e-07,
      "loss": 0.0182,
      "step": 286
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.026786599700167415,
      "grad_norm": 0.12859350442886353,
      "learning_rate": 2e-07,
      "loss": 0.0228,
      "step": 287
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.026879932800168,
      "grad_norm": 0.12045934796333313,
      "learning_rate": 2e-07,
      "loss": 0.0165,
      "step": 288
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.026973265900168582,
      "grad_norm": 0.11929840594530106,
      "learning_rate": 2e-07,
      "loss": 0.0396,
      "step": 289
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.027066599000169167,
      "grad_norm": 0.11182188987731934,
      "learning_rate": 2e-07,
      "loss": 0.0102,
      "step": 290
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02715993210016975,
      "grad_norm": 0.11942093819379807,
      "learning_rate": 2e-07,
      "loss": -0.0245,
      "step": 291
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.027253265200170334,
      "grad_norm": 0.116909921169281,
      "learning_rate": 2e-07,
      "loss": 0.0272,
      "step": 292
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.027346598300170916,
      "grad_norm": 0.12012860178947449,
      "learning_rate": 2e-07,
      "loss": 0.0483,
      "step": 293
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0274399314001715,
      "grad_norm": 0.11081382632255554,
      "learning_rate": 2e-07,
      "loss": 0.0389,
      "step": 294
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.027533264500172083,
      "grad_norm": 0.11655379086732864,
      "learning_rate": 2e-07,
      "loss": 0.0133,
      "step": 295
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.027626597600172668,
      "grad_norm": 0.12386627495288849,
      "learning_rate": 2e-07,
      "loss": 0.011,
      "step": 296
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02771993070017325,
      "grad_norm": 0.13282302021980286,
      "learning_rate": 2e-07,
      "loss": 0.0175,
      "step": 297
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02781326380017383,
      "grad_norm": 0.12305509299039841,
      "learning_rate": 2e-07,
      "loss": 0.0232,
      "step": 298
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.027906596900174416,
      "grad_norm": 0.12764370441436768,
      "learning_rate": 2e-07,
      "loss": 0.0465,
      "step": 299
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.027999930000174998,
      "grad_norm": 0.11132095009088516,
      "learning_rate": 2e-07,
      "loss": 0.0498,
      "step": 300
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.028093263100175583,
      "grad_norm": 0.11686699837446213,
      "learning_rate": 2e-07,
      "loss": 0.0014,
      "step": 301
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.028186596200176165,
      "grad_norm": 0.1156989187002182,
      "learning_rate": 2e-07,
      "loss": 0.031,
      "step": 302
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02827992930017675,
      "grad_norm": 0.12015429139137268,
      "learning_rate": 2e-07,
      "loss": 0.092,
      "step": 303
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.028373262400177332,
      "grad_norm": 0.10886608809232712,
      "learning_rate": 2e-07,
      "loss": 0.038,
      "step": 304
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.028466595500177917,
      "grad_norm": 0.12010408937931061,
      "learning_rate": 2e-07,
      "loss": -0.0338,
      "step": 305
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0285599286001785,
      "grad_norm": 0.12743006646633148,
      "learning_rate": 2e-07,
      "loss": 0.065,
      "step": 306
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.028653261700179084,
      "grad_norm": 0.11614226549863815,
      "learning_rate": 2e-07,
      "loss": 0.0538,
      "step": 307
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.028746594800179666,
      "grad_norm": 0.11234788596630096,
      "learning_rate": 2e-07,
      "loss": 0.0313,
      "step": 308
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02883992790018025,
      "grad_norm": 0.11347447335720062,
      "learning_rate": 2e-07,
      "loss": 0.015,
      "step": 309
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.028933261000180833,
      "grad_norm": 0.1314207762479782,
      "learning_rate": 2e-07,
      "loss": -0.0089,
      "step": 310
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.029026594100181415,
      "grad_norm": 0.12442416697740555,
      "learning_rate": 2e-07,
      "loss": 0.0896,
      "step": 311
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.029119927200182,
      "grad_norm": 0.11614236235618591,
      "learning_rate": 2e-07,
      "loss": 0.0082,
      "step": 312
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02921326030018258,
      "grad_norm": 0.12755386531352997,
      "learning_rate": 2e-07,
      "loss": 0.0157,
      "step": 313
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.029306593400183167,
      "grad_norm": 0.11465509980916977,
      "learning_rate": 2e-07,
      "loss": -0.0032,
      "step": 314
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.02939992650018375,
      "grad_norm": 0.12995555996894836,
      "learning_rate": 2e-07,
      "loss": 0.0251,
      "step": 315
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.029493259600184334,
      "grad_norm": 0.13707414269447327,
      "learning_rate": 2e-07,
      "loss": 0.0313,
      "step": 316
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.029586592700184915,
      "grad_norm": 0.12843500077724457,
      "learning_rate": 2e-07,
      "loss": 0.0607,
      "step": 317
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0296799258001855,
      "grad_norm": 0.11542893201112747,
      "learning_rate": 2e-07,
      "loss": -0.0149,
      "step": 318
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.029773258900186082,
      "grad_norm": 0.12541645765304565,
      "learning_rate": 2e-07,
      "loss": 0.0162,
      "step": 319
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.029866592000186667,
      "grad_norm": 0.1312170922756195,
      "learning_rate": 2e-07,
      "loss": 0.014,
      "step": 320
    },
    {
      "clip_ratio/high_max": 0.007659014401724562,
      "clip_ratio/high_mean": 0.003439755688305013,
      "clip_ratio/low_mean": 0.002222136659838725,
      "clip_ratio/low_min": 0.0002078174802591093,
      "clip_ratio/region_mean": 0.005661892442731187,
      "completions/clipped_ratio": 0.014404296875,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4090.0,
      "completions/mean_length": 609.98681640625,
      "completions/mean_terminated_length": 559.039306640625,
      "completions/min_length": 20.0,
      "completions/min_terminated_length": 20.0,
      "epoch": 0.02995992510018725,
      "grad_norm": 0.11453958600759506,
      "learning_rate": 2e-07,
      "loss": 0.0271,
      "num_tokens": 242896201.0,
      "reward": 0.5606515407562256,
      "reward_std": 0.20513592660427094,
      "rewards/simpleverify_reward/mean": 0.5606514811515808,
      "rewards/simpleverify_reward/std": 0.49631205201148987,
      "step": 321
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.030053258200187834,
      "grad_norm": 0.11788046360015869,
      "learning_rate": 2e-07,
      "loss": -0.0174,
      "step": 322
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.030146591300188416,
      "grad_norm": 0.12439335137605667,
      "learning_rate": 2e-07,
      "loss": 0.0568,
      "step": 323
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.030239924400188998,
      "grad_norm": 0.11043991148471832,
      "learning_rate": 2e-07,
      "loss": 0.0301,
      "step": 324
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.030333257500189583,
      "grad_norm": 0.15738534927368164,
      "learning_rate": 2e-07,
      "loss": 0.0921,
      "step": 325
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.030426590600190165,
      "grad_norm": 0.1203802227973938,
      "learning_rate": 2e-07,
      "loss": 0.0169,
      "step": 326
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03051992370019075,
      "grad_norm": 0.12494948506355286,
      "learning_rate": 2e-07,
      "loss": 0.0357,
      "step": 327
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.030613256800191332,
      "grad_norm": 0.11916864663362503,
      "learning_rate": 2e-07,
      "loss": -0.0414,
      "step": 328
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.030706589900191917,
      "grad_norm": 0.11881279945373535,
      "learning_rate": 2e-07,
      "loss": 0.0315,
      "step": 329
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0307999230001925,
      "grad_norm": 0.12076196819543839,
      "learning_rate": 2e-07,
      "loss": 0.0072,
      "step": 330
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.030893256100193084,
      "grad_norm": 0.12143860012292862,
      "learning_rate": 2e-07,
      "loss": 0.0246,
      "step": 331
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.030986589200193666,
      "grad_norm": 0.12284310907125473,
      "learning_rate": 2e-07,
      "loss": -0.0249,
      "step": 332
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03107992230019425,
      "grad_norm": 0.11416839808225632,
      "learning_rate": 2e-07,
      "loss": 0.0132,
      "step": 333
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.031173255400194833,
      "grad_norm": 0.1261385828256607,
      "learning_rate": 2e-07,
      "loss": 0.0864,
      "step": 334
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03126658850019542,
      "grad_norm": 0.11621243506669998,
      "learning_rate": 2e-07,
      "loss": -0.0043,
      "step": 335
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.031359921600196,
      "grad_norm": 0.11200471967458725,
      "learning_rate": 2e-07,
      "loss": 0.008,
      "step": 336
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03145325470019658,
      "grad_norm": 0.13593006134033203,
      "learning_rate": 2e-07,
      "loss": -0.0065,
      "step": 337
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.031546587800197166,
      "grad_norm": 0.12017598003149033,
      "learning_rate": 2e-07,
      "loss": 0.0413,
      "step": 338
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03163992090019775,
      "grad_norm": 0.11766191571950912,
      "learning_rate": 2e-07,
      "loss": 0.0022,
      "step": 339
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03173325400019833,
      "grad_norm": 0.12265715003013611,
      "learning_rate": 2e-07,
      "loss": -0.0016,
      "step": 340
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.031826587100198915,
      "grad_norm": 0.1262984275817871,
      "learning_rate": 2e-07,
      "loss": 0.0071,
      "step": 341
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0319199202001995,
      "grad_norm": 0.11907317489385605,
      "learning_rate": 2e-07,
      "loss": 0.0478,
      "step": 342
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.032013253300200085,
      "grad_norm": 0.1633232980966568,
      "learning_rate": 2e-07,
      "loss": 0.0285,
      "step": 343
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.032106586400200664,
      "grad_norm": 0.11573426425457001,
      "learning_rate": 2e-07,
      "loss": 0.0139,
      "step": 344
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03219991950020125,
      "grad_norm": 0.11459071934223175,
      "learning_rate": 2e-07,
      "loss": 0.0555,
      "step": 345
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.032293252600201834,
      "grad_norm": 0.11378493905067444,
      "learning_rate": 2e-07,
      "loss": 0.0129,
      "step": 346
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03238658570020242,
      "grad_norm": 0.11712933331727982,
      "learning_rate": 2e-07,
      "loss": 0.0479,
      "step": 347
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.032479918800203,
      "grad_norm": 0.12378206849098206,
      "learning_rate": 2e-07,
      "loss": -0.0291,
      "step": 348
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03257325190020358,
      "grad_norm": 0.11774018406867981,
      "learning_rate": 2e-07,
      "loss": 0.0424,
      "step": 349
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03266658500020417,
      "grad_norm": 0.11160732060670853,
      "learning_rate": 2e-07,
      "loss": 0.0686,
      "step": 350
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.032759918100204746,
      "grad_norm": 0.1186843290925026,
      "learning_rate": 2e-07,
      "loss": 0.0547,
      "step": 351
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03285325120020533,
      "grad_norm": 0.1090741753578186,
      "learning_rate": 2e-07,
      "loss": 0.0303,
      "step": 352
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03294658430020592,
      "grad_norm": 0.11925852298736572,
      "learning_rate": 2e-07,
      "loss": 0.0263,
      "step": 353
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0330399174002065,
      "grad_norm": 0.1273898333311081,
      "learning_rate": 2e-07,
      "loss": 0.0313,
      "step": 354
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03313325050020708,
      "grad_norm": 0.128389373421669,
      "learning_rate": 2e-07,
      "loss": 0.011,
      "step": 355
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.033226583600207665,
      "grad_norm": 0.12090560793876648,
      "learning_rate": 2e-07,
      "loss": 0.0393,
      "step": 356
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03331991670020825,
      "grad_norm": 0.12877219915390015,
      "learning_rate": 2e-07,
      "loss": 0.0315,
      "step": 357
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.033413249800208836,
      "grad_norm": 0.13702577352523804,
      "learning_rate": 2e-07,
      "loss": 0.0422,
      "step": 358
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.033506582900209414,
      "grad_norm": 0.12415044009685516,
      "learning_rate": 2e-07,
      "loss": 0.023,
      "step": 359
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03359991600021,
      "grad_norm": 0.11190974712371826,
      "learning_rate": 2e-07,
      "loss": 0.0343,
      "step": 360
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.033693249100210584,
      "grad_norm": 0.11857246607542038,
      "learning_rate": 2e-07,
      "loss": 0.0174,
      "step": 361
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03378658220021117,
      "grad_norm": 0.11898130923509598,
      "learning_rate": 2e-07,
      "loss": 0.0709,
      "step": 362
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03387991530021175,
      "grad_norm": 0.12032792717218399,
      "learning_rate": 2e-07,
      "loss": 0.051,
      "step": 363
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03397324840021233,
      "grad_norm": 0.11882678419351578,
      "learning_rate": 2e-07,
      "loss": 0.0545,
      "step": 364
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03406658150021292,
      "grad_norm": 0.11008977144956589,
      "learning_rate": 2e-07,
      "loss": 0.048,
      "step": 365
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.034159914600213497,
      "grad_norm": 0.11396142095327377,
      "learning_rate": 2e-07,
      "loss": 0.0365,
      "step": 366
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03425324770021408,
      "grad_norm": 0.1285429745912552,
      "learning_rate": 2e-07,
      "loss": 0.0125,
      "step": 367
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03434658080021467,
      "grad_norm": 0.12839408218860626,
      "learning_rate": 2e-07,
      "loss": 0.0241,
      "step": 368
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03443991390021525,
      "grad_norm": 0.11845999211072922,
      "learning_rate": 2e-07,
      "loss": 0.0288,
      "step": 369
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03453324700021583,
      "grad_norm": 0.10861969739198685,
      "learning_rate": 2e-07,
      "loss": 0.0435,
      "step": 370
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.034626580100216416,
      "grad_norm": 0.11874452233314514,
      "learning_rate": 2e-07,
      "loss": 0.0309,
      "step": 371
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.034719913200217,
      "grad_norm": 0.1427195817232132,
      "learning_rate": 2e-07,
      "loss": 0.043,
      "step": 372
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.034813246300217586,
      "grad_norm": 0.12500236928462982,
      "learning_rate": 2e-07,
      "loss": 0.0071,
      "step": 373
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.034906579400218164,
      "grad_norm": 0.12108566612005234,
      "learning_rate": 2e-07,
      "loss": 0.0399,
      "step": 374
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03499991250021875,
      "grad_norm": 0.11117847263813019,
      "learning_rate": 2e-07,
      "loss": 0.0311,
      "step": 375
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.035093245600219335,
      "grad_norm": 0.11412937939167023,
      "learning_rate": 2e-07,
      "loss": 0.0649,
      "step": 376
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03518657870021991,
      "grad_norm": 0.11501997709274292,
      "learning_rate": 2e-07,
      "loss": 0.0279,
      "step": 377
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0352799118002205,
      "grad_norm": 0.10797776281833649,
      "learning_rate": 2e-07,
      "loss": 0.0364,
      "step": 378
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03537324490022108,
      "grad_norm": 0.1246580109000206,
      "learning_rate": 2e-07,
      "loss": -0.0078,
      "step": 379
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03546657800022167,
      "grad_norm": 0.11960603296756744,
      "learning_rate": 2e-07,
      "loss": 0.0679,
      "step": 380
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03555991110022225,
      "grad_norm": 0.12690047919750214,
      "learning_rate": 2e-07,
      "loss": -0.0046,
      "step": 381
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03565324420022283,
      "grad_norm": 0.1212008073925972,
      "learning_rate": 2e-07,
      "loss": 0.0458,
      "step": 382
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03574657730022342,
      "grad_norm": 0.11976614594459534,
      "learning_rate": 2e-07,
      "loss": 0.0094,
      "step": 383
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.035839910400224,
      "grad_norm": 0.12451492995023727,
      "learning_rate": 2e-07,
      "loss": 0.0013,
      "step": 384
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015363420758928603,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4074.0,
      "completions/mean_length": 616.3922729492188,
      "completions/mean_terminated_length": 562.0994873046875,
      "completions/min_length": 36.0,
      "completions/min_terminated_length": 36.0,
      "epoch": 0.03593324350022458,
      "grad_norm": 0.11171372979879379,
      "learning_rate": 2e-07,
      "loss": 0.0331,
      "num_tokens": 283969903.0,
      "reward": 0.5599365234375,
      "reward_std": 0.2054661214351654,
      "rewards/simpleverify_reward/mean": 0.5599365234375,
      "rewards/simpleverify_reward/std": 0.4963989853858948,
      "step": 385
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.036026576600225166,
      "grad_norm": 0.11575938761234283,
      "learning_rate": 2e-07,
      "loss": 0.0334,
      "step": 386
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03611990970022575,
      "grad_norm": 0.12197114527225494,
      "learning_rate": 2e-07,
      "loss": 0.0425,
      "step": 387
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03621324280022633,
      "grad_norm": 0.1176382303237915,
      "learning_rate": 2e-07,
      "loss": 0.0689,
      "step": 388
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.036306575900226914,
      "grad_norm": 0.12390800565481186,
      "learning_rate": 2e-07,
      "loss": 0.0209,
      "step": 389
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0363999090002275,
      "grad_norm": 0.11948312073945999,
      "learning_rate": 2e-07,
      "loss": 0.0087,
      "step": 390
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.036493242100228085,
      "grad_norm": 0.13085635006427765,
      "learning_rate": 2e-07,
      "loss": 0.0376,
      "step": 391
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03658657520022866,
      "grad_norm": 0.119266077876091,
      "learning_rate": 2e-07,
      "loss": -0.0046,
      "step": 392
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03667990830022925,
      "grad_norm": 0.11295431852340698,
      "learning_rate": 2e-07,
      "loss": 0.0317,
      "step": 393
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.036773241400229834,
      "grad_norm": 0.11514849960803986,
      "learning_rate": 2e-07,
      "loss": 0.057,
      "step": 394
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03686657450023042,
      "grad_norm": 0.11986764520406723,
      "learning_rate": 2e-07,
      "loss": 0.0096,
      "step": 395
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.036959907600231,
      "grad_norm": 0.1266181766986847,
      "learning_rate": 2e-07,
      "loss": 0.0448,
      "step": 396
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03705324070023158,
      "grad_norm": 0.13177832961082458,
      "learning_rate": 2e-07,
      "loss": 0.0036,
      "step": 397
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03714657380023217,
      "grad_norm": 0.11497318744659424,
      "learning_rate": 2e-07,
      "loss": -0.0068,
      "step": 398
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03723990690023275,
      "grad_norm": 0.13660897314548492,
      "learning_rate": 2e-07,
      "loss": 0.0214,
      "step": 399
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03733324000023333,
      "grad_norm": 0.1235608235001564,
      "learning_rate": 2e-07,
      "loss": 0.0762,
      "step": 400
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.037426573100233916,
      "grad_norm": 0.10542558878660202,
      "learning_rate": 2e-07,
      "loss": 0.0184,
      "step": 401
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0375199062002345,
      "grad_norm": 0.11293075233697891,
      "learning_rate": 2e-07,
      "loss": 0.0687,
      "step": 402
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03761323930023508,
      "grad_norm": 0.11791659146547318,
      "learning_rate": 2e-07,
      "loss": -0.0036,
      "step": 403
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.037706572400235665,
      "grad_norm": 0.11492893844842911,
      "learning_rate": 2e-07,
      "loss": -0.0362,
      "step": 404
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03779990550023625,
      "grad_norm": 0.12748655676841736,
      "learning_rate": 2e-07,
      "loss": 0.0239,
      "step": 405
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.037893238600236835,
      "grad_norm": 0.11961893737316132,
      "learning_rate": 2e-07,
      "loss": 0.0365,
      "step": 406
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03798657170023741,
      "grad_norm": 0.11468523740768433,
      "learning_rate": 2e-07,
      "loss": 0.0155,
      "step": 407
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.038079904800238,
      "grad_norm": 0.1257142424583435,
      "learning_rate": 2e-07,
      "loss": 0.0635,
      "step": 408
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.038173237900238584,
      "grad_norm": 0.12088916450738907,
      "learning_rate": 2e-07,
      "loss": -0.0164,
      "step": 409
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03826657100023917,
      "grad_norm": 0.1149362325668335,
      "learning_rate": 2e-07,
      "loss": 0.0101,
      "step": 410
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03835990410023975,
      "grad_norm": 0.11720459908246994,
      "learning_rate": 2e-07,
      "loss": 0.0574,
      "step": 411
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03845323720024033,
      "grad_norm": 0.11584582179784775,
      "learning_rate": 2e-07,
      "loss": 0.0412,
      "step": 412
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03854657030024092,
      "grad_norm": 0.12471593916416168,
      "learning_rate": 2e-07,
      "loss": 0.0329,
      "step": 413
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.038639903400241496,
      "grad_norm": 0.11379550397396088,
      "learning_rate": 2e-07,
      "loss": 0.044,
      "step": 414
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03873323650024208,
      "grad_norm": 0.11682140827178955,
      "learning_rate": 2e-07,
      "loss": -0.018,
      "step": 415
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.038826569600242666,
      "grad_norm": 0.1138271912932396,
      "learning_rate": 2e-07,
      "loss": 0.015,
      "step": 416
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03891990270024325,
      "grad_norm": 0.1217326819896698,
      "learning_rate": 2e-07,
      "loss": -0.0015,
      "step": 417
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03901323580024383,
      "grad_norm": 0.12329524010419846,
      "learning_rate": 2e-07,
      "loss": -0.0005,
      "step": 418
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.039106568900244415,
      "grad_norm": 0.12177414447069168,
      "learning_rate": 2e-07,
      "loss": 0.0277,
      "step": 419
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.039199902000245,
      "grad_norm": 0.12002681940793991,
      "learning_rate": 2e-07,
      "loss": 0.0399,
      "step": 420
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.039293235100245585,
      "grad_norm": 0.12456163763999939,
      "learning_rate": 2e-07,
      "loss": 0.0044,
      "step": 421
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.039386568200246164,
      "grad_norm": 0.11519532650709152,
      "learning_rate": 2e-07,
      "loss": 0.028,
      "step": 422
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03947990130024675,
      "grad_norm": 0.11130990833044052,
      "learning_rate": 2e-07,
      "loss": 0.0393,
      "step": 423
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.039573234400247334,
      "grad_norm": 0.11588078737258911,
      "learning_rate": 2e-07,
      "loss": 0.0315,
      "step": 424
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03966656750024792,
      "grad_norm": 0.11550088226795197,
      "learning_rate": 2e-07,
      "loss": 0.058,
      "step": 425
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0397599006002485,
      "grad_norm": 0.12595392763614655,
      "learning_rate": 2e-07,
      "loss": 0.0118,
      "step": 426
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03985323370024908,
      "grad_norm": 0.12484121322631836,
      "learning_rate": 2e-07,
      "loss": 0.0389,
      "step": 427
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.03994656680024967,
      "grad_norm": 0.11769852787256241,
      "learning_rate": 2e-07,
      "loss": 0.035,
      "step": 428
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.040039899900250246,
      "grad_norm": 0.11621078103780746,
      "learning_rate": 2e-07,
      "loss": 0.0064,
      "step": 429
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04013323300025083,
      "grad_norm": 0.11027857661247253,
      "learning_rate": 2e-07,
      "loss": 0.0507,
      "step": 430
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04022656610025142,
      "grad_norm": 0.11220718175172806,
      "learning_rate": 2e-07,
      "loss": 0.0227,
      "step": 431
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.040319899200252,
      "grad_norm": 0.12183266133069992,
      "learning_rate": 2e-07,
      "loss": 0.053,
      "step": 432
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04041323230025258,
      "grad_norm": 0.12204910814762115,
      "learning_rate": 2e-07,
      "loss": 0.019,
      "step": 433
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.040506565400253165,
      "grad_norm": 0.12495990842580795,
      "learning_rate": 2e-07,
      "loss": 0.0201,
      "step": 434
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04059989850025375,
      "grad_norm": 0.12262093275785446,
      "learning_rate": 2e-07,
      "loss": 0.0418,
      "step": 435
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.040693231600254336,
      "grad_norm": 0.12112521380186081,
      "learning_rate": 2e-07,
      "loss": 0.0599,
      "step": 436
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.040786564700254914,
      "grad_norm": 0.11415773630142212,
      "learning_rate": 2e-07,
      "loss": 0.0425,
      "step": 437
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0408798978002555,
      "grad_norm": 0.11473115533590317,
      "learning_rate": 2e-07,
      "loss": 0.0512,
      "step": 438
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.040973230900256084,
      "grad_norm": 0.11164270341396332,
      "learning_rate": 2e-07,
      "loss": 0.0042,
      "step": 439
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04106656400025666,
      "grad_norm": 0.1157250851392746,
      "learning_rate": 2e-07,
      "loss": 0.0145,
      "step": 440
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04115989710025725,
      "grad_norm": 0.13490960001945496,
      "learning_rate": 2e-07,
      "loss": 0.0383,
      "step": 441
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04125323020025783,
      "grad_norm": 0.11610034853219986,
      "learning_rate": 2e-07,
      "loss": 0.0263,
      "step": 442
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04134656330025842,
      "grad_norm": 0.1263304501771927,
      "learning_rate": 2e-07,
      "loss": 0.0536,
      "step": 443
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.041439896400258996,
      "grad_norm": 0.1265595555305481,
      "learning_rate": 2e-07,
      "loss": 0.0161,
      "step": 444
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04153322950025958,
      "grad_norm": 0.12974615395069122,
      "learning_rate": 2e-07,
      "loss": 0.0123,
      "step": 445
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04162656260026017,
      "grad_norm": 0.11630167812108994,
      "learning_rate": 2e-07,
      "loss": 0.024,
      "step": 446
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04171989570026075,
      "grad_norm": 0.12233869731426239,
      "learning_rate": 2e-07,
      "loss": 0.0134,
      "step": 447
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04181322880026133,
      "grad_norm": 0.12354002892971039,
      "learning_rate": 2e-07,
      "loss": 0.035,
      "step": 448
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.014369419642857095,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4084.0,
      "completions/mean_length": 609.72900390625,
      "completions/mean_terminated_length": 558.9029541015625,
      "completions/min_length": 94.0,
      "completions/min_terminated_length": 94.0,
      "epoch": 0.041906561900261916,
      "grad_norm": 0.12005385011434555,
      "learning_rate": 2e-07,
      "loss": 0.039,
      "num_tokens": 324656226.0,
      "reward": 0.5673479437828064,
      "reward_std": 0.2025642991065979,
      "rewards/simpleverify_reward/mean": 0.5673479437828064,
      "rewards/simpleverify_reward/std": 0.4954477846622467,
      "step": 449
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0419998950002625,
      "grad_norm": 0.111904576420784,
      "learning_rate": 2e-07,
      "loss": 0.0713,
      "step": 450
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.042093228100263086,
      "grad_norm": 0.11571154743432999,
      "learning_rate": 2e-07,
      "loss": 0.0067,
      "step": 451
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.042186561200263664,
      "grad_norm": 0.12770473957061768,
      "learning_rate": 2e-07,
      "loss": 0.0297,
      "step": 452
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04227989430026425,
      "grad_norm": 0.10647688060998917,
      "learning_rate": 2e-07,
      "loss": 0.0154,
      "step": 453
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.042373227400264835,
      "grad_norm": 0.11231386661529541,
      "learning_rate": 2e-07,
      "loss": 0.0165,
      "step": 454
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04246656050026541,
      "grad_norm": 0.11723629385232925,
      "learning_rate": 2e-07,
      "loss": 0.0302,
      "step": 455
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.042559893600266,
      "grad_norm": 0.13797250390052795,
      "learning_rate": 2e-07,
      "loss": -0.013,
      "step": 456
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04265322670026658,
      "grad_norm": 0.1287396252155304,
      "learning_rate": 2e-07,
      "loss": 0.0386,
      "step": 457
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04274655980026717,
      "grad_norm": 0.11675171554088593,
      "learning_rate": 2e-07,
      "loss": 0.0429,
      "step": 458
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04283989290026775,
      "grad_norm": 0.1157287135720253,
      "learning_rate": 2e-07,
      "loss": 0.0297,
      "step": 459
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04293322600026833,
      "grad_norm": 0.11516328901052475,
      "learning_rate": 2e-07,
      "loss": -0.0116,
      "step": 460
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04302655910026892,
      "grad_norm": 0.11692577600479126,
      "learning_rate": 2e-07,
      "loss": 0.0477,
      "step": 461
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0431198922002695,
      "grad_norm": 0.12244383245706558,
      "learning_rate": 2e-07,
      "loss": 0.0036,
      "step": 462
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04321322530027008,
      "grad_norm": 0.1258409023284912,
      "learning_rate": 2e-07,
      "loss": 0.055,
      "step": 463
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.043306558400270666,
      "grad_norm": 0.12354937195777893,
      "learning_rate": 2e-07,
      "loss": 0.0226,
      "step": 464
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04339989150027125,
      "grad_norm": 0.11941950768232346,
      "learning_rate": 2e-07,
      "loss": 0.02,
      "step": 465
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04349322460027183,
      "grad_norm": 0.12062560766935349,
      "learning_rate": 2e-07,
      "loss": 0.0144,
      "step": 466
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.043586557700272414,
      "grad_norm": 0.11342823505401611,
      "learning_rate": 2e-07,
      "loss": 0.0512,
      "step": 467
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.043679890800273,
      "grad_norm": 0.1262807548046112,
      "learning_rate": 2e-07,
      "loss": -0.0059,
      "step": 468
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.043773223900273585,
      "grad_norm": 0.11998648941516876,
      "learning_rate": 2e-07,
      "loss": 0.055,
      "step": 469
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04386655700027416,
      "grad_norm": 0.11615382134914398,
      "learning_rate": 2e-07,
      "loss": -0.0117,
      "step": 470
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04395989010027475,
      "grad_norm": 0.13024789094924927,
      "learning_rate": 2e-07,
      "loss": -0.0026,
      "step": 471
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.044053223200275334,
      "grad_norm": 0.12381007522344589,
      "learning_rate": 2e-07,
      "loss": 0.0091,
      "step": 472
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04414655630027592,
      "grad_norm": 0.12598036229610443,
      "learning_rate": 2e-07,
      "loss": 0.0025,
      "step": 473
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0442398894002765,
      "grad_norm": 0.12260746210813522,
      "learning_rate": 2e-07,
      "loss": 0.0382,
      "step": 474
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04433322250027708,
      "grad_norm": 0.12364880740642548,
      "learning_rate": 2e-07,
      "loss": 0.062,
      "step": 475
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04442655560027767,
      "grad_norm": 0.11674712598323822,
      "learning_rate": 2e-07,
      "loss": 0.0135,
      "step": 476
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04451988870027825,
      "grad_norm": 0.13197995722293854,
      "learning_rate": 2e-07,
      "loss": 0.0884,
      "step": 477
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04461322180027883,
      "grad_norm": 0.13497941195964813,
      "learning_rate": 2e-07,
      "loss": 0.0441,
      "step": 478
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.044706554900279416,
      "grad_norm": 0.11924539506435394,
      "learning_rate": 2e-07,
      "loss": 0.0474,
      "step": 479
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04479988800028,
      "grad_norm": 0.103081613779068,
      "learning_rate": 2e-07,
      "loss": 0.0364,
      "step": 480
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04489322110028058,
      "grad_norm": 0.13224312663078308,
      "learning_rate": 2e-07,
      "loss": 0.0208,
      "step": 481
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.044986554200281165,
      "grad_norm": 0.11899323016405106,
      "learning_rate": 2e-07,
      "loss": 0.0197,
      "step": 482
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04507988730028175,
      "grad_norm": 0.12379021942615509,
      "learning_rate": 2e-07,
      "loss": 0.005,
      "step": 483
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.045173220400282335,
      "grad_norm": 0.11254280060529709,
      "learning_rate": 2e-07,
      "loss": 0.047,
      "step": 484
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04526655350028291,
      "grad_norm": 0.1226729080080986,
      "learning_rate": 2e-07,
      "loss": 0.0211,
      "step": 485
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0453598866002835,
      "grad_norm": 0.12863241136074066,
      "learning_rate": 2e-07,
      "loss": -0.0305,
      "step": 486
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.045453219700284084,
      "grad_norm": 0.12958678603172302,
      "learning_rate": 2e-07,
      "loss": 0.0206,
      "step": 487
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04554655280028467,
      "grad_norm": 0.12602970004081726,
      "learning_rate": 2e-07,
      "loss": 0.0223,
      "step": 488
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04563988590028525,
      "grad_norm": 0.11636984348297119,
      "learning_rate": 2e-07,
      "loss": 0.0206,
      "step": 489
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04573321900028583,
      "grad_norm": 0.11772146075963974,
      "learning_rate": 2e-07,
      "loss": 0.0514,
      "step": 490
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04582655210028642,
      "grad_norm": 0.1220271959900856,
      "learning_rate": 2e-07,
      "loss": -0.0046,
      "step": 491
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.045919885200286996,
      "grad_norm": 0.11762391030788422,
      "learning_rate": 2e-07,
      "loss": 0.041,
      "step": 492
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04601321830028758,
      "grad_norm": 0.11399685591459274,
      "learning_rate": 2e-07,
      "loss": 0.0506,
      "step": 493
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.046106551400288166,
      "grad_norm": 0.12811368703842163,
      "learning_rate": 2e-07,
      "loss": 0.0331,
      "step": 494
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04619988450028875,
      "grad_norm": 0.11819914728403091,
      "learning_rate": 2e-07,
      "loss": 0.0667,
      "step": 495
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04629321760028933,
      "grad_norm": 0.12240269035100937,
      "learning_rate": 2e-07,
      "loss": 0.0431,
      "step": 496
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.046386550700289915,
      "grad_norm": 0.11342491954565048,
      "learning_rate": 2e-07,
      "loss": 0.0047,
      "step": 497
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0464798838002905,
      "grad_norm": 0.12665295600891113,
      "learning_rate": 2e-07,
      "loss": 0.0138,
      "step": 498
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.046573216900291085,
      "grad_norm": 0.11746233701705933,
      "learning_rate": 2e-07,
      "loss": 0.031,
      "step": 499
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.046666550000291664,
      "grad_norm": 0.11802034080028534,
      "learning_rate": 2e-07,
      "loss": 0.0471,
      "step": 500
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04675988310029225,
      "grad_norm": 0.17467068135738373,
      "learning_rate": 2e-07,
      "loss": 0.0101,
      "step": 501
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.046853216200292834,
      "grad_norm": 0.1267123967409134,
      "learning_rate": 2e-07,
      "loss": 0.0123,
      "step": 502
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04694654930029342,
      "grad_norm": 0.12186948955059052,
      "learning_rate": 2e-07,
      "loss": 0.0087,
      "step": 503
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.047039882400294,
      "grad_norm": 0.11872043460607529,
      "learning_rate": 2e-07,
      "loss": 0.0029,
      "step": 504
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04713321550029458,
      "grad_norm": 0.10396856814622879,
      "learning_rate": 2e-07,
      "loss": -0.0218,
      "step": 505
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04722654860029517,
      "grad_norm": 0.11877740174531937,
      "learning_rate": 2e-07,
      "loss": 0.0688,
      "step": 506
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.047319881700295746,
      "grad_norm": 0.11273297667503357,
      "learning_rate": 2e-07,
      "loss": 0.0157,
      "step": 507
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04741321480029633,
      "grad_norm": 0.12132605165243149,
      "learning_rate": 2e-07,
      "loss": 0.057,
      "step": 508
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04750654790029692,
      "grad_norm": 0.11349218338727951,
      "learning_rate": 2e-07,
      "loss": 0.0336,
      "step": 509
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0475998810002975,
      "grad_norm": 0.12423590570688248,
      "learning_rate": 2e-07,
      "loss": 0.0168,
      "step": 510
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04769321410029808,
      "grad_norm": 0.12998363375663757,
      "learning_rate": 2e-07,
      "loss": 0.014,
      "step": 511
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.047786547200298665,
      "grad_norm": 0.1233840137720108,
      "learning_rate": 2e-07,
      "loss": 0.0127,
      "step": 512
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.014753069196428603,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4093.0,
      "completions/mean_length": 616.5823974609375,
      "completions/mean_terminated_length": 564.4816284179688,
      "completions/min_length": 14.0,
      "completions/min_terminated_length": 14.0,
      "epoch": 0.04787988030029925,
      "grad_norm": 0.12010251730680466,
      "learning_rate": 2e-07,
      "loss": 0.0267,
      "num_tokens": 365726406.0,
      "reward": 0.5629185438156128,
      "reward_std": 0.20161347091197968,
      "rewards/simpleverify_reward/mean": 0.5629185438156128,
      "rewards/simpleverify_reward/std": 0.496029794216156,
      "step": 513
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.047973213400299836,
      "grad_norm": 0.12401369959115982,
      "learning_rate": 2e-07,
      "loss": 0.0836,
      "step": 514
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.048066546500300414,
      "grad_norm": 0.12161466479301453,
      "learning_rate": 2e-07,
      "loss": 0.0406,
      "step": 515
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.048159879600301,
      "grad_norm": 0.11125585436820984,
      "learning_rate": 2e-07,
      "loss": 0.0133,
      "step": 516
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.048253212700301584,
      "grad_norm": 0.11706260591745377,
      "learning_rate": 2e-07,
      "loss": -0.0356,
      "step": 517
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04834654580030216,
      "grad_norm": 0.12694256007671356,
      "learning_rate": 2e-07,
      "loss": -0.0134,
      "step": 518
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04843987890030275,
      "grad_norm": 0.12294843792915344,
      "learning_rate": 2e-07,
      "loss": 0.0201,
      "step": 519
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04853321200030333,
      "grad_norm": 0.1124502494931221,
      "learning_rate": 2e-07,
      "loss": 0.0809,
      "step": 520
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04862654510030392,
      "grad_norm": 0.112555593252182,
      "learning_rate": 2e-07,
      "loss": 0.0489,
      "step": 521
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.048719878200304496,
      "grad_norm": 0.11587313562631607,
      "learning_rate": 2e-07,
      "loss": 0.0318,
      "step": 522
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04881321130030508,
      "grad_norm": 0.12432841211557388,
      "learning_rate": 2e-07,
      "loss": 0.0464,
      "step": 523
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04890654440030567,
      "grad_norm": 0.11995571106672287,
      "learning_rate": 2e-07,
      "loss": 0.0179,
      "step": 524
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04899987750030625,
      "grad_norm": 0.12222591787576675,
      "learning_rate": 2e-07,
      "loss": 0.0345,
      "step": 525
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04909321060030683,
      "grad_norm": 0.12418544292449951,
      "learning_rate": 2e-07,
      "loss": -0.0131,
      "step": 526
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.049186543700307415,
      "grad_norm": 0.12256152182817459,
      "learning_rate": 2e-07,
      "loss": 0.0498,
      "step": 527
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.049279876800308,
      "grad_norm": 0.11545681953430176,
      "learning_rate": 2e-07,
      "loss": 0.0307,
      "step": 528
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.049373209900308586,
      "grad_norm": 0.11925394088029861,
      "learning_rate": 2e-07,
      "loss": 0.0234,
      "step": 529
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.049466543000309164,
      "grad_norm": 0.12285978347063065,
      "learning_rate": 2e-07,
      "loss": -0.0017,
      "step": 530
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04955987610030975,
      "grad_norm": 0.12846523523330688,
      "learning_rate": 2e-07,
      "loss": 0.0268,
      "step": 531
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.049653209200310335,
      "grad_norm": 0.12642474472522736,
      "learning_rate": 2e-07,
      "loss": 0.0386,
      "step": 532
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04974654230031091,
      "grad_norm": 0.11340638250112534,
      "learning_rate": 2e-07,
      "loss": 0.0533,
      "step": 533
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0498398754003115,
      "grad_norm": 0.1254989504814148,
      "learning_rate": 2e-07,
      "loss": -0.0036,
      "step": 534
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.04993320850031208,
      "grad_norm": 0.11798431724309921,
      "learning_rate": 2e-07,
      "loss": 0.0462,
      "step": 535
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05002654160031267,
      "grad_norm": 0.120912104845047,
      "learning_rate": 2e-07,
      "loss": 0.0146,
      "step": 536
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05011987470031325,
      "grad_norm": 0.12695574760437012,
      "learning_rate": 2e-07,
      "loss": 0.0651,
      "step": 537
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05021320780031383,
      "grad_norm": 0.11621187627315521,
      "learning_rate": 2e-07,
      "loss": 0.0308,
      "step": 538
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05030654090031442,
      "grad_norm": 0.11845751851797104,
      "learning_rate": 2e-07,
      "loss": 0.0305,
      "step": 539
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.050399874000315,
      "grad_norm": 0.11774702370166779,
      "learning_rate": 2e-07,
      "loss": 0.0046,
      "step": 540
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05049320710031558,
      "grad_norm": 0.11731205135583878,
      "learning_rate": 2e-07,
      "loss": 0.0199,
      "step": 541
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.050586540200316166,
      "grad_norm": 0.12474443763494492,
      "learning_rate": 2e-07,
      "loss": 0.017,
      "step": 542
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05067987330031675,
      "grad_norm": 0.12901581823825836,
      "learning_rate": 2e-07,
      "loss": 0.0712,
      "step": 543
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05077320640031733,
      "grad_norm": 0.12137272953987122,
      "learning_rate": 2e-07,
      "loss": -0.0208,
      "step": 544
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.050866539500317914,
      "grad_norm": 0.11454559862613678,
      "learning_rate": 2e-07,
      "loss": 0.0446,
      "step": 545
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0509598726003185,
      "grad_norm": 0.11357691884040833,
      "learning_rate": 2e-07,
      "loss": -0.0066,
      "step": 546
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.051053205700319085,
      "grad_norm": 0.12128932029008865,
      "learning_rate": 2e-07,
      "loss": 0.0085,
      "step": 547
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05114653880031966,
      "grad_norm": 0.12682893872261047,
      "learning_rate": 2e-07,
      "loss": 0.0452,
      "step": 548
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05123987190032025,
      "grad_norm": 0.11873123049736023,
      "learning_rate": 2e-07,
      "loss": 0.0378,
      "step": 549
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05133320500032083,
      "grad_norm": 0.11859776824712753,
      "learning_rate": 2e-07,
      "loss": -0.0072,
      "step": 550
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05142653810032142,
      "grad_norm": 0.13037604093551636,
      "learning_rate": 2e-07,
      "loss": -0.0015,
      "step": 551
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.051519871200322,
      "grad_norm": 0.12165237218141556,
      "learning_rate": 2e-07,
      "loss": 0.0473,
      "step": 552
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05161320430032258,
      "grad_norm": 0.10922785103321075,
      "learning_rate": 2e-07,
      "loss": 0.0226,
      "step": 553
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05170653740032317,
      "grad_norm": 0.10690672695636749,
      "learning_rate": 2e-07,
      "loss": 0.0224,
      "step": 554
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05179987050032375,
      "grad_norm": 0.11512351036071777,
      "learning_rate": 2e-07,
      "loss": 0.004,
      "step": 555
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05189320360032433,
      "grad_norm": 0.1353617012500763,
      "learning_rate": 2e-07,
      "loss": -0.0119,
      "step": 556
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.051986536700324916,
      "grad_norm": 0.21192969381809235,
      "learning_rate": 2e-07,
      "loss": 0.0529,
      "step": 557
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0520798698003255,
      "grad_norm": 0.11621417105197906,
      "learning_rate": 2e-07,
      "loss": 0.0025,
      "step": 558
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05217320290032608,
      "grad_norm": 0.1171375960111618,
      "learning_rate": 2e-07,
      "loss": 0.0645,
      "step": 559
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.052266536000326665,
      "grad_norm": 0.12459057569503784,
      "learning_rate": 2e-07,
      "loss": 0.0123,
      "step": 560
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05235986910032725,
      "grad_norm": 0.1224704161286354,
      "learning_rate": 2e-07,
      "loss": 0.0502,
      "step": 561
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.052453202200327835,
      "grad_norm": 0.12132956087589264,
      "learning_rate": 2e-07,
      "loss": 0.0058,
      "step": 562
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05254653530032841,
      "grad_norm": 0.1108483150601387,
      "learning_rate": 2e-07,
      "loss": 0.0575,
      "step": 563
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.052639868400329,
      "grad_norm": 0.12339775264263153,
      "learning_rate": 2e-07,
      "loss": -0.012,
      "step": 564
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.052733201500329584,
      "grad_norm": 0.10971226543188095,
      "learning_rate": 2e-07,
      "loss": 0.04,
      "step": 565
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05282653460033017,
      "grad_norm": 0.12245997786521912,
      "learning_rate": 2e-07,
      "loss": 0.0452,
      "step": 566
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05291986770033075,
      "grad_norm": 0.12251701205968857,
      "learning_rate": 2e-07,
      "loss": 0.0105,
      "step": 567
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05301320080033133,
      "grad_norm": 0.11216636747121811,
      "learning_rate": 2e-07,
      "loss": 0.024,
      "step": 568
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05310653390033192,
      "grad_norm": 0.1176990494132042,
      "learning_rate": 2e-07,
      "loss": 0.0197,
      "step": 569
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.053199867000332496,
      "grad_norm": 0.12697085738182068,
      "learning_rate": 2e-07,
      "loss": 0.0239,
      "step": 570
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05329320010033308,
      "grad_norm": 0.12195334583520889,
      "learning_rate": 2e-07,
      "loss": 0.0298,
      "step": 571
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.053386533200333666,
      "grad_norm": 0.11798093467950821,
      "learning_rate": 2e-07,
      "loss": 0.06,
      "step": 572
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05347986630033425,
      "grad_norm": 0.11503004282712936,
      "learning_rate": 2e-07,
      "loss": -0.0118,
      "step": 573
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05357319940033483,
      "grad_norm": 0.11584347486495972,
      "learning_rate": 2e-07,
      "loss": 0.0595,
      "step": 574
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.053666532500335415,
      "grad_norm": 0.12356378883123398,
      "learning_rate": 2e-07,
      "loss": 0.0522,
      "step": 575
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.053759865600336,
      "grad_norm": 0.12094777077436447,
      "learning_rate": 2e-07,
      "loss": 0.0167,
      "step": 576
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.015869140625,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4094.0,
      "completions/mean_length": 625.546630859375,
      "completions/mean_terminated_length": 569.58544921875,
      "completions/min_length": 82.0,
      "completions/min_terminated_length": 82.0,
      "epoch": 0.053853198700336585,
      "grad_norm": 0.12044382840394974,
      "learning_rate": 2e-07,
      "loss": 0.0311,
      "num_tokens": 407349056.0,
      "reward": 0.5599365234375,
      "reward_std": 0.20234021544456482,
      "rewards/simpleverify_reward/mean": 0.5599365234375,
      "rewards/simpleverify_reward/std": 0.4963989853858948,
      "step": 577
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.053946531800337164,
      "grad_norm": 0.11953672766685486,
      "learning_rate": 2e-07,
      "loss": 0.0715,
      "step": 578
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05403986490033775,
      "grad_norm": 0.12127688527107239,
      "learning_rate": 2e-07,
      "loss": 0.0208,
      "step": 579
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.054133198000338334,
      "grad_norm": 0.12586002051830292,
      "learning_rate": 2e-07,
      "loss": -0.038,
      "step": 580
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05422653110033892,
      "grad_norm": 0.10979132354259491,
      "learning_rate": 2e-07,
      "loss": 0.0477,
      "step": 581
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0543198642003395,
      "grad_norm": 0.12793982028961182,
      "learning_rate": 2e-07,
      "loss": 0.0262,
      "step": 582
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05441319730034008,
      "grad_norm": 0.1122364029288292,
      "learning_rate": 2e-07,
      "loss": 0.0315,
      "step": 583
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05450653040034067,
      "grad_norm": 0.1163412481546402,
      "learning_rate": 2e-07,
      "loss": 0.0696,
      "step": 584
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.054599863500341246,
      "grad_norm": 0.11988294869661331,
      "learning_rate": 2e-07,
      "loss": 0.0072,
      "step": 585
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05469319660034183,
      "grad_norm": 0.11390503495931625,
      "learning_rate": 2e-07,
      "loss": -0.0012,
      "step": 586
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.054786529700342416,
      "grad_norm": 0.12465047091245651,
      "learning_rate": 2e-07,
      "loss": 0.0085,
      "step": 587
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.054879862800343,
      "grad_norm": 0.11787502467632294,
      "learning_rate": 2e-07,
      "loss": 0.0265,
      "step": 588
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05497319590034358,
      "grad_norm": 0.12637466192245483,
      "learning_rate": 2e-07,
      "loss": 0.0607,
      "step": 589
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.055066529000344165,
      "grad_norm": 0.11292491853237152,
      "learning_rate": 2e-07,
      "loss": 0.0227,
      "step": 590
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05515986210034475,
      "grad_norm": 0.12993048131465912,
      "learning_rate": 2e-07,
      "loss": 0.0229,
      "step": 591
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.055253195200345336,
      "grad_norm": 0.10965694487094879,
      "learning_rate": 2e-07,
      "loss": 0.0405,
      "step": 592
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.055346528300345914,
      "grad_norm": 0.12179083377122879,
      "learning_rate": 2e-07,
      "loss": 0.0402,
      "step": 593
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0554398614003465,
      "grad_norm": 0.1250147670507431,
      "learning_rate": 2e-07,
      "loss": -0.0507,
      "step": 594
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.055533194500347084,
      "grad_norm": 0.11787499487400055,
      "learning_rate": 2e-07,
      "loss": -0.0143,
      "step": 595
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05562652760034766,
      "grad_norm": 0.12058252096176147,
      "learning_rate": 2e-07,
      "loss": 0.0578,
      "step": 596
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05571986070034825,
      "grad_norm": 0.11161614954471588,
      "learning_rate": 2e-07,
      "loss": 0.0465,
      "step": 597
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05581319380034883,
      "grad_norm": 0.11848585307598114,
      "learning_rate": 2e-07,
      "loss": 0.0747,
      "step": 598
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05590652690034942,
      "grad_norm": 0.1354919672012329,
      "learning_rate": 2e-07,
      "loss": 0.0425,
      "step": 599
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.055999860000349996,
      "grad_norm": 0.12854740023612976,
      "learning_rate": 2e-07,
      "loss": 0.0058,
      "step": 600
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05609319310035058,
      "grad_norm": 0.1257350742816925,
      "learning_rate": 2e-07,
      "loss": 0.0417,
      "step": 601
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05618652620035117,
      "grad_norm": 0.11217089742422104,
      "learning_rate": 2e-07,
      "loss": 0.0166,
      "step": 602
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05627985930035175,
      "grad_norm": 0.12380943447351456,
      "learning_rate": 2e-07,
      "loss": 0.0433,
      "step": 603
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05637319240035233,
      "grad_norm": 0.12836982309818268,
      "learning_rate": 2e-07,
      "loss": 0.0084,
      "step": 604
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.056466525500352915,
      "grad_norm": 0.12369835376739502,
      "learning_rate": 2e-07,
      "loss": 0.0228,
      "step": 605
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0565598586003535,
      "grad_norm": 0.12817060947418213,
      "learning_rate": 2e-07,
      "loss": 0.0585,
      "step": 606
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.056653191700354086,
      "grad_norm": 0.1351020783185959,
      "learning_rate": 2e-07,
      "loss": 0.0436,
      "step": 607
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.056746524800354664,
      "grad_norm": 0.13817666471004486,
      "learning_rate": 2e-07,
      "loss": 0.0267,
      "step": 608
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05683985790035525,
      "grad_norm": 0.1242447942495346,
      "learning_rate": 2e-07,
      "loss": 0.0016,
      "step": 609
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.056933191000355834,
      "grad_norm": 0.12094396352767944,
      "learning_rate": 2e-07,
      "loss": 0.043,
      "step": 610
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05702652410035641,
      "grad_norm": 0.11039764434099197,
      "learning_rate": 2e-07,
      "loss": 0.0177,
      "step": 611
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.057119857200357,
      "grad_norm": 0.12412373721599579,
      "learning_rate": 2e-07,
      "loss": 0.0507,
      "step": 612
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05721319030035758,
      "grad_norm": 0.12555938959121704,
      "learning_rate": 2e-07,
      "loss": 0.0393,
      "step": 613
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05730652340035817,
      "grad_norm": 0.11187642067670822,
      "learning_rate": 2e-07,
      "loss": 0.0236,
      "step": 614
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05739985650035875,
      "grad_norm": 0.11063692718744278,
      "learning_rate": 2e-07,
      "loss": 0.0386,
      "step": 615
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05749318960035933,
      "grad_norm": 0.12599587440490723,
      "learning_rate": 2e-07,
      "loss": 0.0117,
      "step": 616
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05758652270035992,
      "grad_norm": 0.12586668133735657,
      "learning_rate": 2e-07,
      "loss": 0.0194,
      "step": 617
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0576798558003605,
      "grad_norm": 0.13283438980579376,
      "learning_rate": 2e-07,
      "loss": 0.0414,
      "step": 618
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05777318890036108,
      "grad_norm": 0.12420965731143951,
      "learning_rate": 2e-07,
      "loss": 0.0465,
      "step": 619
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.057866522000361666,
      "grad_norm": 0.1160416528582573,
      "learning_rate": 2e-07,
      "loss": 0.0362,
      "step": 620
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05795985510036225,
      "grad_norm": 0.12362019717693329,
      "learning_rate": 2e-07,
      "loss": 0.0924,
      "step": 621
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05805318820036283,
      "grad_norm": 0.12439006567001343,
      "learning_rate": 2e-07,
      "loss": 0.0085,
      "step": 622
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.058146521300363414,
      "grad_norm": 0.13101525604724884,
      "learning_rate": 2e-07,
      "loss": 0.0178,
      "step": 623
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.058239854400364,
      "grad_norm": 0.1352616548538208,
      "learning_rate": 2e-07,
      "loss": 0.0304,
      "step": 624
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.058333187500364585,
      "grad_norm": 0.12140699476003647,
      "learning_rate": 2e-07,
      "loss": 0.0126,
      "step": 625
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05842652060036516,
      "grad_norm": 0.12117926776409149,
      "learning_rate": 2e-07,
      "loss": 0.0585,
      "step": 626
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05851985370036575,
      "grad_norm": 0.13625787198543549,
      "learning_rate": 2e-07,
      "loss": -0.0058,
      "step": 627
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05861318680036633,
      "grad_norm": 0.1238526925444603,
      "learning_rate": 2e-07,
      "loss": 0.0615,
      "step": 628
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05870651990036692,
      "grad_norm": 0.12498913705348969,
      "learning_rate": 2e-07,
      "loss": 0.0354,
      "step": 629
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0587998530003675,
      "grad_norm": 0.10788042098283768,
      "learning_rate": 2e-07,
      "loss": 0.0406,
      "step": 630
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05889318610036808,
      "grad_norm": 0.09922719746828079,
      "learning_rate": 2e-07,
      "loss": 0.0742,
      "step": 631
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05898651920036867,
      "grad_norm": 0.12002391368150711,
      "learning_rate": 2e-07,
      "loss": 0.0098,
      "step": 632
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05907985230036925,
      "grad_norm": 0.13086529076099396,
      "learning_rate": 2e-07,
      "loss": 0.0533,
      "step": 633
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05917318540036983,
      "grad_norm": 0.1261654645204544,
      "learning_rate": 2e-07,
      "loss": 0.0205,
      "step": 634
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.059266518500370416,
      "grad_norm": 0.12553098797798157,
      "learning_rate": 2e-07,
      "loss": -0.0116,
      "step": 635
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.059359851600371,
      "grad_norm": 0.12851980328559875,
      "learning_rate": 2e-07,
      "loss": 0.0237,
      "step": 636
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05945318470037158,
      "grad_norm": 0.12041840702295303,
      "learning_rate": 2e-07,
      "loss": -0.0246,
      "step": 637
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.059546517800372165,
      "grad_norm": 0.11303598433732986,
      "learning_rate": 2e-07,
      "loss": -0.0066,
      "step": 638
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.05963985090037275,
      "grad_norm": 0.13044410943984985,
      "learning_rate": 2e-07,
      "loss": 0.0573,
      "step": 639
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.059733184000373335,
      "grad_norm": 0.11876334995031357,
      "learning_rate": 2e-07,
      "loss": 0.036,
      "step": 640
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.014107840401785698,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4096.0,
      "completions/mean_length": 615.0515747070312,
      "completions/mean_terminated_length": 565.2401733398438,
      "completions/min_length": 89.0,
      "completions/min_terminated_length": 89.0,
      "epoch": 0.05982651710037391,
      "grad_norm": 0.12729071080684662,
      "learning_rate": 2e-07,
      "loss": 0.0173,
      "num_tokens": 448368123.0,
      "reward": 0.5781599283218384,
      "reward_std": 0.19839850068092346,
      "rewards/simpleverify_reward/mean": 0.5781598687171936,
      "rewards/simpleverify_reward/std": 0.49385756254196167,
      "step": 641
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0599198502003745,
      "grad_norm": 0.13411562144756317,
      "learning_rate": 2e-07,
      "loss": 0.0245,
      "step": 642
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.060013183300375084,
      "grad_norm": 0.129892036318779,
      "learning_rate": 2e-07,
      "loss": 0.0098,
      "step": 643
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06010651640037567,
      "grad_norm": 0.11809277534484863,
      "learning_rate": 2e-07,
      "loss": 0.0443,
      "step": 644
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06019984950037625,
      "grad_norm": 0.10522212833166122,
      "learning_rate": 2e-07,
      "loss": 0.0345,
      "step": 645
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06029318260037683,
      "grad_norm": 0.13501420617103577,
      "learning_rate": 2e-07,
      "loss": 0.0616,
      "step": 646
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06038651570037742,
      "grad_norm": 0.12211109697818756,
      "learning_rate": 2e-07,
      "loss": 0.0129,
      "step": 647
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.060479848800377996,
      "grad_norm": 0.11742499470710754,
      "learning_rate": 2e-07,
      "loss": 0.0271,
      "step": 648
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06057318190037858,
      "grad_norm": 0.11299612373113632,
      "learning_rate": 2e-07,
      "loss": 0.0164,
      "step": 649
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.060666515000379166,
      "grad_norm": 0.1275217980146408,
      "learning_rate": 2e-07,
      "loss": 0.0818,
      "step": 650
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06075984810037975,
      "grad_norm": 0.1271250993013382,
      "learning_rate": 2e-07,
      "loss": 0.0342,
      "step": 651
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06085318120038033,
      "grad_norm": 0.13171058893203735,
      "learning_rate": 2e-07,
      "loss": -0.0025,
      "step": 652
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.060946514300380915,
      "grad_norm": 0.12337896227836609,
      "learning_rate": 2e-07,
      "loss": 0.015,
      "step": 653
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0610398474003815,
      "grad_norm": 0.12394645810127258,
      "learning_rate": 2e-07,
      "loss": 0.0436,
      "step": 654
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.061133180500382085,
      "grad_norm": 0.11018452793359756,
      "learning_rate": 2e-07,
      "loss": 0.079,
      "step": 655
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.061226513600382663,
      "grad_norm": 0.11741793155670166,
      "learning_rate": 2e-07,
      "loss": 0.0177,
      "step": 656
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06131984670038325,
      "grad_norm": 0.11924643069505692,
      "learning_rate": 2e-07,
      "loss": 0.0028,
      "step": 657
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.061413179800383834,
      "grad_norm": 0.11948590725660324,
      "learning_rate": 2e-07,
      "loss": -0.019,
      "step": 658
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06150651290038442,
      "grad_norm": 0.11440945416688919,
      "learning_rate": 2e-07,
      "loss": 0.0107,
      "step": 659
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.061599846000385,
      "grad_norm": 0.12413817644119263,
      "learning_rate": 2e-07,
      "loss": 0.0293,
      "step": 660
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06169317910038558,
      "grad_norm": 0.12349940836429596,
      "learning_rate": 2e-07,
      "loss": 0.0244,
      "step": 661
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06178651220038617,
      "grad_norm": 0.13031722605228424,
      "learning_rate": 2e-07,
      "loss": -0.003,
      "step": 662
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.061879845300386746,
      "grad_norm": 0.11165805160999298,
      "learning_rate": 2e-07,
      "loss": 0.0323,
      "step": 663
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06197317840038733,
      "grad_norm": 0.11943699419498444,
      "learning_rate": 2e-07,
      "loss": 0.0322,
      "step": 664
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.062066511500387916,
      "grad_norm": 0.10237414389848709,
      "learning_rate": 2e-07,
      "loss": 0.0357,
      "step": 665
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0621598446003885,
      "grad_norm": 0.14724992215633392,
      "learning_rate": 2e-07,
      "loss": 0.0206,
      "step": 666
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06225317770038908,
      "grad_norm": 0.11832781881093979,
      "learning_rate": 2e-07,
      "loss": 0.0615,
      "step": 667
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.062346510800389665,
      "grad_norm": 0.12036070972681046,
      "learning_rate": 2e-07,
      "loss": -0.0115,
      "step": 668
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06243984390039025,
      "grad_norm": 0.12463673949241638,
      "learning_rate": 2e-07,
      "loss": 0.0009,
      "step": 669
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06253317700039084,
      "grad_norm": 0.12530124187469482,
      "learning_rate": 2e-07,
      "loss": 0.0237,
      "step": 670
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06262651010039141,
      "grad_norm": 0.13599103689193726,
      "learning_rate": 2e-07,
      "loss": 0.0792,
      "step": 671
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.062719843200392,
      "grad_norm": 0.12094172090291977,
      "learning_rate": 2e-07,
      "loss": -0.0046,
      "step": 672
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06281317630039258,
      "grad_norm": 0.11020464450120926,
      "learning_rate": 2e-07,
      "loss": 0.0168,
      "step": 673
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06290650940039316,
      "grad_norm": 0.12409074604511261,
      "learning_rate": 2e-07,
      "loss": 0.0158,
      "step": 674
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06299984250039375,
      "grad_norm": 0.12593719363212585,
      "learning_rate": 2e-07,
      "loss": 0.0537,
      "step": 675
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06309317560039433,
      "grad_norm": 0.1168484091758728,
      "learning_rate": 2e-07,
      "loss": 0.0272,
      "step": 676
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06318650870039491,
      "grad_norm": 0.11330760270357132,
      "learning_rate": 2e-07,
      "loss": 0.0236,
      "step": 677
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0632798418003955,
      "grad_norm": 0.125793918967247,
      "learning_rate": 2e-07,
      "loss": 0.0541,
      "step": 678
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06337317490039608,
      "grad_norm": 0.12242653220891953,
      "learning_rate": 2e-07,
      "loss": -0.0508,
      "step": 679
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06346650800039666,
      "grad_norm": 0.1316857784986496,
      "learning_rate": 2e-07,
      "loss": 0.0169,
      "step": 680
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06355984110039725,
      "grad_norm": 0.1167135164141655,
      "learning_rate": 2e-07,
      "loss": 0.0087,
      "step": 681
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06365317420039783,
      "grad_norm": 0.11635976284742355,
      "learning_rate": 2e-07,
      "loss": 0.0711,
      "step": 682
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06374650730039842,
      "grad_norm": 0.10801029950380325,
      "learning_rate": 2e-07,
      "loss": 0.06,
      "step": 683
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.063839840400399,
      "grad_norm": 0.12761594355106354,
      "learning_rate": 2e-07,
      "loss": 0.0146,
      "step": 684
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06393317350039958,
      "grad_norm": 0.12739582359790802,
      "learning_rate": 2e-07,
      "loss": 0.0169,
      "step": 685
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06402650660040017,
      "grad_norm": 0.12135310471057892,
      "learning_rate": 2e-07,
      "loss": -0.0238,
      "step": 686
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06411983970040075,
      "grad_norm": 0.12800715863704681,
      "learning_rate": 2e-07,
      "loss": 0.0015,
      "step": 687
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06421317280040133,
      "grad_norm": 0.11485518515110016,
      "learning_rate": 2e-07,
      "loss": 0.0185,
      "step": 688
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06430650590040192,
      "grad_norm": 0.10507942736148834,
      "learning_rate": 2e-07,
      "loss": -0.0023,
      "step": 689
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0643998390004025,
      "grad_norm": 0.11504718661308289,
      "learning_rate": 2e-07,
      "loss": 0.0298,
      "step": 690
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06449317210040308,
      "grad_norm": 0.14104816317558289,
      "learning_rate": 2e-07,
      "loss": 0.0274,
      "step": 691
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06458650520040367,
      "grad_norm": 0.11345266550779343,
      "learning_rate": 2e-07,
      "loss": 0.0339,
      "step": 692
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06467983830040425,
      "grad_norm": 0.12505804002285004,
      "learning_rate": 2e-07,
      "loss": 0.0422,
      "step": 693
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06477317140040484,
      "grad_norm": 0.12232425808906555,
      "learning_rate": 2e-07,
      "loss": 0.0532,
      "step": 694
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06486650450040542,
      "grad_norm": 0.10989726334810257,
      "learning_rate": 2e-07,
      "loss": 0.0472,
      "step": 695
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.064959837600406,
      "grad_norm": 0.12593911588191986,
      "learning_rate": 2e-07,
      "loss": 0.0156,
      "step": 696
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06505317070040659,
      "grad_norm": 0.12165261805057526,
      "learning_rate": 2e-07,
      "loss": 0.0283,
      "step": 697
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06514650380040717,
      "grad_norm": 0.11726663261651993,
      "learning_rate": 2e-07,
      "loss": 0.0509,
      "step": 698
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06523983690040774,
      "grad_norm": 0.11849267780780792,
      "learning_rate": 2e-07,
      "loss": 0.024,
      "step": 699
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06533317000040834,
      "grad_norm": 0.11234334856271744,
      "learning_rate": 2e-07,
      "loss": 0.018,
      "step": 700
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06542650310040891,
      "grad_norm": 0.12003482133150101,
      "learning_rate": 2e-07,
      "loss": -0.0038,
      "step": 701
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06551983620040949,
      "grad_norm": 0.12619444727897644,
      "learning_rate": 2e-07,
      "loss": 0.022,
      "step": 702
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06561316930041008,
      "grad_norm": 0.12507031857967377,
      "learning_rate": 2e-07,
      "loss": 0.0494,
      "step": 703
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06570650240041066,
      "grad_norm": 0.12343506515026093,
      "learning_rate": 2e-07,
      "loss": 0.0418,
      "step": 704
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.014299665178571397,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4091.0,
      "completions/mean_length": 620.75537109375,
      "completions/mean_terminated_length": 570.339599609375,
      "completions/min_length": 85.0,
      "completions/min_terminated_length": 85.0,
      "epoch": 0.06579983550041126,
      "grad_norm": 0.12528972327709198,
      "learning_rate": 2e-07,
      "loss": -0.0102,
      "num_tokens": 489691551.0,
      "reward": 0.5700858235359192,
      "reward_std": 0.19754192233085632,
      "rewards/simpleverify_reward/mean": 0.5700858235359192,
      "rewards/simpleverify_reward/std": 0.4950679540634155,
      "step": 705
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06589316860041183,
      "grad_norm": 0.1104116141796112,
      "learning_rate": 2e-07,
      "loss": 0.0105,
      "step": 706
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06598650170041241,
      "grad_norm": 0.12788976728916168,
      "learning_rate": 2e-07,
      "loss": -0.0125,
      "step": 707
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.066079834800413,
      "grad_norm": 0.1163189634680748,
      "learning_rate": 2e-07,
      "loss": 0.0251,
      "step": 708
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06617316790041358,
      "grad_norm": 0.11647484451532364,
      "learning_rate": 2e-07,
      "loss": 0.0366,
      "step": 709
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06626650100041416,
      "grad_norm": 0.12332244962453842,
      "learning_rate": 2e-07,
      "loss": 0.0463,
      "step": 710
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06635983410041475,
      "grad_norm": 0.11935242265462875,
      "learning_rate": 2e-07,
      "loss": 0.0586,
      "step": 711
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06645316720041533,
      "grad_norm": 0.10812757164239883,
      "learning_rate": 2e-07,
      "loss": 0.0196,
      "step": 712
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06654650030041591,
      "grad_norm": 0.1112031415104866,
      "learning_rate": 2e-07,
      "loss": 0.054,
      "step": 713
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0666398334004165,
      "grad_norm": 0.13041889667510986,
      "learning_rate": 2e-07,
      "loss": 0.021,
      "step": 714
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06673316650041708,
      "grad_norm": 0.12210946530103683,
      "learning_rate": 2e-07,
      "loss": 0.0012,
      "step": 715
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06682649960041767,
      "grad_norm": 0.12628982961177826,
      "learning_rate": 2e-07,
      "loss": 0.0403,
      "step": 716
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06691983270041825,
      "grad_norm": 0.12646602094173431,
      "learning_rate": 2e-07,
      "loss": 0.0136,
      "step": 717
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06701316580041883,
      "grad_norm": 0.1331997960805893,
      "learning_rate": 2e-07,
      "loss": 0.0267,
      "step": 718
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06710649890041942,
      "grad_norm": 0.13744772970676422,
      "learning_rate": 2e-07,
      "loss": 0.0239,
      "step": 719
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06719983200042,
      "grad_norm": 0.12118663638830185,
      "learning_rate": 2e-07,
      "loss": 0.0253,
      "step": 720
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06729316510042058,
      "grad_norm": 0.1154777854681015,
      "learning_rate": 2e-07,
      "loss": 0.0117,
      "step": 721
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06738649820042117,
      "grad_norm": 0.11925779283046722,
      "learning_rate": 2e-07,
      "loss": 0.0025,
      "step": 722
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06747983130042175,
      "grad_norm": 0.11497519910335541,
      "learning_rate": 2e-07,
      "loss": 0.0385,
      "step": 723
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06757316440042234,
      "grad_norm": 0.10929624736309052,
      "learning_rate": 2e-07,
      "loss": 0.0265,
      "step": 724
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06766649750042292,
      "grad_norm": 0.12020684033632278,
      "learning_rate": 2e-07,
      "loss": 0.052,
      "step": 725
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0677598306004235,
      "grad_norm": 0.1540072113275528,
      "learning_rate": 2e-07,
      "loss": 0.0138,
      "step": 726
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06785316370042409,
      "grad_norm": 0.10292994976043701,
      "learning_rate": 2e-07,
      "loss": -0.0066,
      "step": 727
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06794649680042467,
      "grad_norm": 0.13004785776138306,
      "learning_rate": 2e-07,
      "loss": 0.0245,
      "step": 728
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06803982990042524,
      "grad_norm": 0.12006964534521103,
      "learning_rate": 2e-07,
      "loss": 0.0026,
      "step": 729
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06813316300042584,
      "grad_norm": 0.11545645445585251,
      "learning_rate": 2e-07,
      "loss": 0.0154,
      "step": 730
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06822649610042641,
      "grad_norm": 0.11766402423381805,
      "learning_rate": 2e-07,
      "loss": 0.0505,
      "step": 731
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06831982920042699,
      "grad_norm": 0.11055393517017365,
      "learning_rate": 2e-07,
      "loss": 0.0176,
      "step": 732
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06841316230042759,
      "grad_norm": 0.11462956666946411,
      "learning_rate": 2e-07,
      "loss": 0.0122,
      "step": 733
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06850649540042816,
      "grad_norm": 0.10686997324228287,
      "learning_rate": 2e-07,
      "loss": 0.0605,
      "step": 734
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06859982850042876,
      "grad_norm": 0.1133904904127121,
      "learning_rate": 2e-07,
      "loss": 0.0305,
      "step": 735
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06869316160042933,
      "grad_norm": 0.1171252503991127,
      "learning_rate": 2e-07,
      "loss": 0.0631,
      "step": 736
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06878649470042991,
      "grad_norm": 0.11815008521080017,
      "learning_rate": 2e-07,
      "loss": 0.0314,
      "step": 737
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0688798278004305,
      "grad_norm": 0.12277720123529434,
      "learning_rate": 2e-07,
      "loss": 0.0232,
      "step": 738
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06897316090043108,
      "grad_norm": 0.12605513632297516,
      "learning_rate": 2e-07,
      "loss": 0.0418,
      "step": 739
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06906649400043166,
      "grad_norm": 0.12493447959423065,
      "learning_rate": 2e-07,
      "loss": 0.0183,
      "step": 740
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06915982710043225,
      "grad_norm": 0.1309005320072174,
      "learning_rate": 2e-07,
      "loss": 0.0358,
      "step": 741
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06925316020043283,
      "grad_norm": 0.11564654856920242,
      "learning_rate": 2e-07,
      "loss": 0.0537,
      "step": 742
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06934649330043341,
      "grad_norm": 0.11595161259174347,
      "learning_rate": 2e-07,
      "loss": -0.0148,
      "step": 743
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.069439826400434,
      "grad_norm": 0.1234142929315567,
      "learning_rate": 2e-07,
      "loss": 0.0345,
      "step": 744
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06953315950043458,
      "grad_norm": 0.12163940072059631,
      "learning_rate": 2e-07,
      "loss": 0.0429,
      "step": 745
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06962649260043517,
      "grad_norm": 0.12371745705604553,
      "learning_rate": 2e-07,
      "loss": 0.0779,
      "step": 746
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06971982570043575,
      "grad_norm": 0.12303513288497925,
      "learning_rate": 2e-07,
      "loss": 0.0324,
      "step": 747
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06981315880043633,
      "grad_norm": 0.1302696019411087,
      "learning_rate": 2e-07,
      "loss": 0.0414,
      "step": 748
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.06990649190043692,
      "grad_norm": 0.11138232797384262,
      "learning_rate": 2e-07,
      "loss": 0.0375,
      "step": 749
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0699998250004375,
      "grad_norm": 0.11783741414546967,
      "learning_rate": 2e-07,
      "loss": 0.0422,
      "step": 750
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07009315810043808,
      "grad_norm": 0.11582102626562119,
      "learning_rate": 2e-07,
      "loss": 0.0314,
      "step": 751
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07018649120043867,
      "grad_norm": 0.11814740300178528,
      "learning_rate": 2e-07,
      "loss": 0.0264,
      "step": 752
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07027982430043925,
      "grad_norm": 0.12032865732908249,
      "learning_rate": 2e-07,
      "loss": 0.027,
      "step": 753
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07037315740043983,
      "grad_norm": 0.12482430040836334,
      "learning_rate": 2e-07,
      "loss": -0.0004,
      "step": 754
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07046649050044042,
      "grad_norm": 0.1225215271115303,
      "learning_rate": 2e-07,
      "loss": 0.0662,
      "step": 755
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.070559823600441,
      "grad_norm": 0.1313469111919403,
      "learning_rate": 2e-07,
      "loss": 0.038,
      "step": 756
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07065315670044159,
      "grad_norm": 0.12063515931367874,
      "learning_rate": 2e-07,
      "loss": -0.0188,
      "step": 757
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07074648980044217,
      "grad_norm": 0.12330294400453568,
      "learning_rate": 2e-07,
      "loss": 0.0557,
      "step": 758
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07083982290044274,
      "grad_norm": 0.11578790843486786,
      "learning_rate": 2e-07,
      "loss": 0.036,
      "step": 759
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07093315600044334,
      "grad_norm": 0.12653319537639618,
      "learning_rate": 2e-07,
      "loss": 0.0478,
      "step": 760
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07102648910044392,
      "grad_norm": 0.11110926419496536,
      "learning_rate": 2e-07,
      "loss": 0.0131,
      "step": 761
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0711198222004445,
      "grad_norm": 0.12657906115055084,
      "learning_rate": 2e-07,
      "loss": -0.0323,
      "step": 762
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07121315530044509,
      "grad_norm": 0.12415838986635208,
      "learning_rate": 2e-07,
      "loss": 0.0426,
      "step": 763
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07130648840044566,
      "grad_norm": 0.12819182872772217,
      "learning_rate": 2e-07,
      "loss": 0.0306,
      "step": 764
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07139982150044624,
      "grad_norm": 0.1336245983839035,
      "learning_rate": 2e-07,
      "loss": -0.043,
      "step": 765
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07149315460044683,
      "grad_norm": 0.12627893686294556,
      "learning_rate": 2e-07,
      "loss": 0.0531,
      "step": 766
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07158648770044741,
      "grad_norm": 0.12721318006515503,
      "learning_rate": 2e-07,
      "loss": -0.0013,
      "step": 767
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.071679820800448,
      "grad_norm": 0.13815297186374664,
      "learning_rate": 2e-07,
      "loss": 0.0259,
      "step": 768
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.013462611607142905,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4095.0,
      "completions/mean_length": 610.3123779296875,
      "completions/mean_terminated_length": 562.74560546875,
      "completions/min_length": 83.0,
      "completions/min_terminated_length": 83.0,
      "epoch": 0.07177315390044858,
      "grad_norm": 0.13740846514701843,
      "learning_rate": 2e-07,
      "loss": 0.0006,
      "num_tokens": 530465705.0,
      "reward": 0.5800955891609192,
      "reward_std": 0.194438174366951,
      "rewards/simpleverify_reward/mean": 0.5800955891609192,
      "rewards/simpleverify_reward/std": 0.493547260761261,
      "step": 769
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07186648700044916,
      "grad_norm": 0.10117055475711823,
      "learning_rate": 2e-07,
      "loss": 0.0509,
      "step": 770
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07195982010044975,
      "grad_norm": 0.12342734634876251,
      "learning_rate": 2e-07,
      "loss": 0.0357,
      "step": 771
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07205315320045033,
      "grad_norm": 0.12916740775108337,
      "learning_rate": 2e-07,
      "loss": 0.034,
      "step": 772
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07214648630045091,
      "grad_norm": 0.11947473883628845,
      "learning_rate": 2e-07,
      "loss": 0.0202,
      "step": 773
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0722398194004515,
      "grad_norm": 0.12576515972614288,
      "learning_rate": 2e-07,
      "loss": 0.0005,
      "step": 774
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07233315250045208,
      "grad_norm": 0.10413672775030136,
      "learning_rate": 2e-07,
      "loss": 0.0016,
      "step": 775
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07242648560045266,
      "grad_norm": 0.1263851374387741,
      "learning_rate": 2e-07,
      "loss": 0.0381,
      "step": 776
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07251981870045325,
      "grad_norm": 0.12243737280368805,
      "learning_rate": 2e-07,
      "loss": -0.0041,
      "step": 777
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07261315180045383,
      "grad_norm": 0.11369921267032623,
      "learning_rate": 2e-07,
      "loss": 0.0585,
      "step": 778
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07270648490045442,
      "grad_norm": 0.12998270988464355,
      "learning_rate": 2e-07,
      "loss": 0.0066,
      "step": 779
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.072799818000455,
      "grad_norm": 0.12854069471359253,
      "learning_rate": 2e-07,
      "loss": 0.0701,
      "step": 780
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07289315110045558,
      "grad_norm": 0.12057509273290634,
      "learning_rate": 2e-07,
      "loss": 0.035,
      "step": 781
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07298648420045617,
      "grad_norm": 0.1271604746580124,
      "learning_rate": 2e-07,
      "loss": 0.0202,
      "step": 782
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07307981730045675,
      "grad_norm": 0.12926845252513885,
      "learning_rate": 2e-07,
      "loss": 0.0012,
      "step": 783
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07317315040045733,
      "grad_norm": 0.12242759019136429,
      "learning_rate": 2e-07,
      "loss": 0.0325,
      "step": 784
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07326648350045792,
      "grad_norm": 0.12419641762971878,
      "learning_rate": 2e-07,
      "loss": -0.0034,
      "step": 785
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0733598166004585,
      "grad_norm": 0.10994675010442734,
      "learning_rate": 2e-07,
      "loss": 0.005,
      "step": 786
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07345314970045909,
      "grad_norm": 0.12860162556171417,
      "learning_rate": 2e-07,
      "loss": -0.009,
      "step": 787
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07354648280045967,
      "grad_norm": 0.12840574979782104,
      "learning_rate": 2e-07,
      "loss": 0.0636,
      "step": 788
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07363981590046025,
      "grad_norm": 0.12296827137470245,
      "learning_rate": 2e-07,
      "loss": 0.0429,
      "step": 789
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07373314900046084,
      "grad_norm": 0.11585507541894913,
      "learning_rate": 2e-07,
      "loss": 0.0262,
      "step": 790
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07382648210046142,
      "grad_norm": 0.12902520596981049,
      "learning_rate": 2e-07,
      "loss": 0.0103,
      "step": 791
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.073919815200462,
      "grad_norm": 0.11695924401283264,
      "learning_rate": 2e-07,
      "loss": 0.0638,
      "step": 792
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07401314830046259,
      "grad_norm": 0.13254743814468384,
      "learning_rate": 2e-07,
      "loss": 0.0255,
      "step": 793
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07410648140046316,
      "grad_norm": 0.11255477368831635,
      "learning_rate": 2e-07,
      "loss": 0.0662,
      "step": 794
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07419981450046374,
      "grad_norm": 0.12763340771198273,
      "learning_rate": 2e-07,
      "loss": 0.0833,
      "step": 795
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07429314760046433,
      "grad_norm": 0.12116031348705292,
      "learning_rate": 2e-07,
      "loss": 0.0073,
      "step": 796
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07438648070046491,
      "grad_norm": 0.11752741038799286,
      "learning_rate": 2e-07,
      "loss": 0.0479,
      "step": 797
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0744798138004655,
      "grad_norm": 0.1262942999601364,
      "learning_rate": 2e-07,
      "loss": -0.0255,
      "step": 798
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07457314690046608,
      "grad_norm": 0.11733046919107437,
      "learning_rate": 2e-07,
      "loss": -0.0031,
      "step": 799
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07466648000046666,
      "grad_norm": 0.12669570744037628,
      "learning_rate": 2e-07,
      "loss": 0.0052,
      "step": 800
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07475981310046725,
      "grad_norm": 0.12124934792518616,
      "learning_rate": 2e-07,
      "loss": 0.0378,
      "step": 801
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07485314620046783,
      "grad_norm": 0.12422847002744675,
      "learning_rate": 2e-07,
      "loss": 0.0019,
      "step": 802
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07494647930046841,
      "grad_norm": 0.11205616593360901,
      "learning_rate": 2e-07,
      "loss": 0.0084,
      "step": 803
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.075039812400469,
      "grad_norm": 0.12309475988149643,
      "learning_rate": 2e-07,
      "loss": 0.0715,
      "step": 804
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07513314550046958,
      "grad_norm": 0.11718499660491943,
      "learning_rate": 2e-07,
      "loss": -0.0219,
      "step": 805
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07522647860047016,
      "grad_norm": 0.11949222534894943,
      "learning_rate": 2e-07,
      "loss": 0.0122,
      "step": 806
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07531981170047075,
      "grad_norm": 0.12406980991363525,
      "learning_rate": 2e-07,
      "loss": 0.0576,
      "step": 807
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07541314480047133,
      "grad_norm": 0.14561213552951813,
      "learning_rate": 2e-07,
      "loss": 0.0315,
      "step": 808
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07550647790047192,
      "grad_norm": 0.11671137809753418,
      "learning_rate": 2e-07,
      "loss": 0.0299,
      "step": 809
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0755998110004725,
      "grad_norm": 0.12124525755643845,
      "learning_rate": 2e-07,
      "loss": -0.0034,
      "step": 810
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07569314410047308,
      "grad_norm": 0.11499109119176865,
      "learning_rate": 2e-07,
      "loss": 0.0489,
      "step": 811
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07578647720047367,
      "grad_norm": 0.11485016345977783,
      "learning_rate": 2e-07,
      "loss": 0.0293,
      "step": 812
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07587981030047425,
      "grad_norm": 0.12557214498519897,
      "learning_rate": 2e-07,
      "loss": 0.0221,
      "step": 813
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07597314340047483,
      "grad_norm": 0.13099895417690277,
      "learning_rate": 2e-07,
      "loss": 0.0095,
      "step": 814
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07606647650047542,
      "grad_norm": 0.12623251974582672,
      "learning_rate": 2e-07,
      "loss": -0.0348,
      "step": 815
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.076159809600476,
      "grad_norm": 0.11864347010850906,
      "learning_rate": 2e-07,
      "loss": 0.0391,
      "step": 816
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07625314270047658,
      "grad_norm": 0.12183661013841629,
      "learning_rate": 2e-07,
      "loss": 0.0529,
      "step": 817
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07634647580047717,
      "grad_norm": 0.132380411028862,
      "learning_rate": 2e-07,
      "loss": 0.012,
      "step": 818
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07643980890047775,
      "grad_norm": 0.11724260449409485,
      "learning_rate": 2e-07,
      "loss": 0.0049,
      "step": 819
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07653314200047834,
      "grad_norm": 0.12289160490036011,
      "learning_rate": 2e-07,
      "loss": 0.0227,
      "step": 820
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07662647510047892,
      "grad_norm": 0.1231679618358612,
      "learning_rate": 2e-07,
      "loss": -0.0061,
      "step": 821
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0767198082004795,
      "grad_norm": 0.12912705540657043,
      "learning_rate": 2e-07,
      "loss": 0.0406,
      "step": 822
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07681314130048009,
      "grad_norm": 0.12607784569263458,
      "learning_rate": 2e-07,
      "loss": 0.0058,
      "step": 823
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07690647440048066,
      "grad_norm": 0.11383119225502014,
      "learning_rate": 2e-07,
      "loss": 0.0103,
      "step": 824
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07699980750048124,
      "grad_norm": 0.11440588533878326,
      "learning_rate": 2e-07,
      "loss": 0.0441,
      "step": 825
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07709314060048184,
      "grad_norm": 0.15354876220226288,
      "learning_rate": 2e-07,
      "loss": 0.1046,
      "step": 826
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07718647370048241,
      "grad_norm": 0.1340963989496231,
      "learning_rate": 2e-07,
      "loss": 0.0378,
      "step": 827
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07727980680048299,
      "grad_norm": 0.12661057710647583,
      "learning_rate": 2e-07,
      "loss": 0.0518,
      "step": 828
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07737313990048358,
      "grad_norm": 0.13152505457401276,
      "learning_rate": 2e-07,
      "loss": 0.0164,
      "step": 829
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07746647300048416,
      "grad_norm": 0.11410197615623474,
      "learning_rate": 2e-07,
      "loss": 0.0579,
      "step": 830
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07755980610048475,
      "grad_norm": 0.11915868520736694,
      "learning_rate": 2e-07,
      "loss": 0.0111,
      "step": 831
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07765313920048533,
      "grad_norm": 0.12689052522182465,
      "learning_rate": 2e-07,
      "loss": 0.0209,
      "step": 832
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.013916015625,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4095.0,
      "completions/mean_length": 620.7677612304688,
      "completions/mean_terminated_length": 571.7238159179688,
      "completions/min_length": 37.0,
      "completions/min_terminated_length": 37.0,
      "epoch": 0.07774647230048591,
      "grad_norm": 0.10465911775827408,
      "learning_rate": 2e-07,
      "loss": 0.0263,
      "num_tokens": 571812769.0,
      "reward": 0.5783517360687256,
      "reward_std": 0.19668343663215637,
      "rewards/simpleverify_reward/mean": 0.5783516764640808,
      "rewards/simpleverify_reward/std": 0.4938271641731262,
      "step": 833
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0778398054004865,
      "grad_norm": 0.10756061226129532,
      "learning_rate": 2e-07,
      "loss": 0.0025,
      "step": 834
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07793313850048708,
      "grad_norm": 0.1274019032716751,
      "learning_rate": 2e-07,
      "loss": 0.0706,
      "step": 835
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07802647160048766,
      "grad_norm": 0.12987348437309265,
      "learning_rate": 2e-07,
      "loss": -0.032,
      "step": 836
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07811980470048825,
      "grad_norm": 0.11911167204380035,
      "learning_rate": 2e-07,
      "loss": 0.0797,
      "step": 837
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07821313780048883,
      "grad_norm": 0.12521375715732574,
      "learning_rate": 2e-07,
      "loss": 0.0468,
      "step": 838
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07830647090048942,
      "grad_norm": 0.11838897317647934,
      "learning_rate": 2e-07,
      "loss": 0.0323,
      "step": 839
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07839980400049,
      "grad_norm": 0.12225809693336487,
      "learning_rate": 2e-07,
      "loss": -0.0205,
      "step": 840
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07849313710049058,
      "grad_norm": 0.11803404986858368,
      "learning_rate": 2e-07,
      "loss": -0.0037,
      "step": 841
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07858647020049117,
      "grad_norm": 0.1223907396197319,
      "learning_rate": 2e-07,
      "loss": -0.0084,
      "step": 842
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07867980330049175,
      "grad_norm": 0.13088178634643555,
      "learning_rate": 2e-07,
      "loss": 0.0267,
      "step": 843
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07877313640049233,
      "grad_norm": 0.1251993626356125,
      "learning_rate": 2e-07,
      "loss": 0.0387,
      "step": 844
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07886646950049292,
      "grad_norm": 0.12215556204319,
      "learning_rate": 2e-07,
      "loss": 0.0091,
      "step": 845
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0789598026004935,
      "grad_norm": 0.128033846616745,
      "learning_rate": 2e-07,
      "loss": 0.0022,
      "step": 846
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07905313570049408,
      "grad_norm": 0.1244906485080719,
      "learning_rate": 2e-07,
      "loss": 0.0296,
      "step": 847
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07914646880049467,
      "grad_norm": 0.13499657809734344,
      "learning_rate": 2e-07,
      "loss": 0.022,
      "step": 848
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07923980190049525,
      "grad_norm": 0.11610082536935806,
      "learning_rate": 2e-07,
      "loss": 0.0274,
      "step": 849
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07933313500049584,
      "grad_norm": 0.12622587382793427,
      "learning_rate": 2e-07,
      "loss": 0.0085,
      "step": 850
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07942646810049642,
      "grad_norm": 0.12759481370449066,
      "learning_rate": 2e-07,
      "loss": 0.0486,
      "step": 851
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.079519801200497,
      "grad_norm": 0.11895918101072311,
      "learning_rate": 2e-07,
      "loss": 0.0291,
      "step": 852
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07961313430049759,
      "grad_norm": 0.12240754812955856,
      "learning_rate": 2e-07,
      "loss": 0.0018,
      "step": 853
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07970646740049817,
      "grad_norm": 0.13033229112625122,
      "learning_rate": 2e-07,
      "loss": 0.0423,
      "step": 854
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07979980050049874,
      "grad_norm": 0.1316843330860138,
      "learning_rate": 2e-07,
      "loss": 0.0134,
      "step": 855
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07989313360049934,
      "grad_norm": 0.12169570475816727,
      "learning_rate": 2e-07,
      "loss": 0.0428,
      "step": 856
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.07998646670049991,
      "grad_norm": 0.12846745550632477,
      "learning_rate": 2e-07,
      "loss": 0.0183,
      "step": 857
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08007979980050049,
      "grad_norm": 0.1295788586139679,
      "learning_rate": 2e-07,
      "loss": 0.0014,
      "step": 858
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08017313290050108,
      "grad_norm": 0.12061750143766403,
      "learning_rate": 2e-07,
      "loss": 0.0294,
      "step": 859
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08026646600050166,
      "grad_norm": 0.12721611559391022,
      "learning_rate": 2e-07,
      "loss": 0.0245,
      "step": 860
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08035979910050225,
      "grad_norm": 0.1279670000076294,
      "learning_rate": 2e-07,
      "loss": 0.0683,
      "step": 861
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08045313220050283,
      "grad_norm": 0.10878216475248337,
      "learning_rate": 2e-07,
      "loss": 0.0068,
      "step": 862
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08054646530050341,
      "grad_norm": 0.1190527155995369,
      "learning_rate": 2e-07,
      "loss": 0.0346,
      "step": 863
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.080639798400504,
      "grad_norm": 0.1198086366057396,
      "learning_rate": 2e-07,
      "loss": -0.0105,
      "step": 864
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08073313150050458,
      "grad_norm": 0.1230640709400177,
      "learning_rate": 2e-07,
      "loss": -0.0022,
      "step": 865
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08082646460050516,
      "grad_norm": 0.12793295085430145,
      "learning_rate": 2e-07,
      "loss": 0.0478,
      "step": 866
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08091979770050575,
      "grad_norm": 0.1162162646651268,
      "learning_rate": 2e-07,
      "loss": 0.0049,
      "step": 867
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08101313080050633,
      "grad_norm": 0.12349235266447067,
      "learning_rate": 2e-07,
      "loss": 0.0411,
      "step": 868
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08110646390050691,
      "grad_norm": 0.12162021547555923,
      "learning_rate": 2e-07,
      "loss": 0.0037,
      "step": 869
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0811997970005075,
      "grad_norm": 0.1159018874168396,
      "learning_rate": 2e-07,
      "loss": 0.0168,
      "step": 870
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08129313010050808,
      "grad_norm": 0.13062871992588043,
      "learning_rate": 2e-07,
      "loss": 0.0835,
      "step": 871
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08138646320050867,
      "grad_norm": 0.11619515717029572,
      "learning_rate": 2e-07,
      "loss": -0.0007,
      "step": 872
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08147979630050925,
      "grad_norm": 0.12824323773384094,
      "learning_rate": 2e-07,
      "loss": 0.0625,
      "step": 873
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08157312940050983,
      "grad_norm": 0.133867546916008,
      "learning_rate": 2e-07,
      "loss": 0.0118,
      "step": 874
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08166646250051042,
      "grad_norm": 0.1203899085521698,
      "learning_rate": 2e-07,
      "loss": 0.0304,
      "step": 875
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.081759795600511,
      "grad_norm": 0.11946319788694382,
      "learning_rate": 2e-07,
      "loss": 0.0062,
      "step": 876
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08185312870051158,
      "grad_norm": 0.12081809341907501,
      "learning_rate": 2e-07,
      "loss": 0.0314,
      "step": 877
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08194646180051217,
      "grad_norm": 0.12231713533401489,
      "learning_rate": 2e-07,
      "loss": -0.0148,
      "step": 878
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08203979490051275,
      "grad_norm": 0.12683087587356567,
      "learning_rate": 2e-07,
      "loss": 0.0375,
      "step": 879
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08213312800051333,
      "grad_norm": 0.117578886449337,
      "learning_rate": 2e-07,
      "loss": 0.0885,
      "step": 880
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08222646110051392,
      "grad_norm": 0.13788649439811707,
      "learning_rate": 2e-07,
      "loss": 0.0422,
      "step": 881
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0823197942005145,
      "grad_norm": 0.1260453760623932,
      "learning_rate": 2e-07,
      "loss": 0.0614,
      "step": 882
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08241312730051509,
      "grad_norm": 0.12063919752836227,
      "learning_rate": 2e-07,
      "loss": 0.0148,
      "step": 883
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08250646040051567,
      "grad_norm": 0.11219226568937302,
      "learning_rate": 2e-07,
      "loss": 0.0447,
      "step": 884
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08259979350051624,
      "grad_norm": 0.12594319880008698,
      "learning_rate": 2e-07,
      "loss": -0.0184,
      "step": 885
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08269312660051684,
      "grad_norm": 0.1485334038734436,
      "learning_rate": 2e-07,
      "loss": 0.0517,
      "step": 886
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08278645970051741,
      "grad_norm": 0.11263196915388107,
      "learning_rate": 2e-07,
      "loss": 0.0144,
      "step": 887
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08287979280051799,
      "grad_norm": 0.12246878445148468,
      "learning_rate": 2e-07,
      "loss": 0.0342,
      "step": 888
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08297312590051859,
      "grad_norm": 0.11180509626865387,
      "learning_rate": 2e-07,
      "loss": 0.026,
      "step": 889
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08306645900051916,
      "grad_norm": 0.12872332334518433,
      "learning_rate": 2e-07,
      "loss": 0.0288,
      "step": 890
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08315979210051976,
      "grad_norm": 0.12065563350915909,
      "learning_rate": 2e-07,
      "loss": 0.0306,
      "step": 891
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08325312520052033,
      "grad_norm": 0.132719486951828,
      "learning_rate": 2e-07,
      "loss": 0.056,
      "step": 892
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08334645830052091,
      "grad_norm": 0.1211332380771637,
      "learning_rate": 2e-07,
      "loss": 0.0119,
      "step": 893
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0834397914005215,
      "grad_norm": 0.13359197974205017,
      "learning_rate": 2e-07,
      "loss": 0.0638,
      "step": 894
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08353312450052208,
      "grad_norm": 0.13562211394309998,
      "learning_rate": 2e-07,
      "loss": 0.0588,
      "step": 895
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08362645760052266,
      "grad_norm": 0.1118512749671936,
      "learning_rate": 2e-07,
      "loss": 0.0426,
      "step": 896
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.013480050223214302,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4088.0,
      "completions/mean_length": 622.762451171875,
      "completions/mean_terminated_length": 575.30322265625,
      "completions/min_length": 24.0,
      "completions/min_terminated_length": 24.0,
      "epoch": 0.08371979070052325,
      "grad_norm": 0.12406393885612488,
      "learning_rate": 2e-07,
      "loss": 0.0198,
      "num_tokens": 613244209.0,
      "reward": 0.577340304851532,
      "reward_std": 0.1989586055278778,
      "rewards/simpleverify_reward/mean": 0.5773402452468872,
      "rewards/simpleverify_reward/std": 0.493986576795578,
      "step": 897
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08381312380052383,
      "grad_norm": 0.11742869764566422,
      "learning_rate": 2e-07,
      "loss": 0.0467,
      "step": 898
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08390645690052441,
      "grad_norm": 0.12272965908050537,
      "learning_rate": 2e-07,
      "loss": 0.0292,
      "step": 899
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.083999790000525,
      "grad_norm": 0.12145863473415375,
      "learning_rate": 2e-07,
      "loss": 0.0382,
      "step": 900
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08409312310052558,
      "grad_norm": 0.12996797263622284,
      "learning_rate": 2e-07,
      "loss": 0.012,
      "step": 901
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08418645620052617,
      "grad_norm": 0.12358734011650085,
      "learning_rate": 2e-07,
      "loss": 0.0215,
      "step": 902
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08427978930052675,
      "grad_norm": 0.1235094666481018,
      "learning_rate": 2e-07,
      "loss": 0.0651,
      "step": 903
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08437312240052733,
      "grad_norm": 0.11705267429351807,
      "learning_rate": 2e-07,
      "loss": 0.022,
      "step": 904
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08446645550052792,
      "grad_norm": 0.12949518859386444,
      "learning_rate": 2e-07,
      "loss": 0.0665,
      "step": 905
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0845597886005285,
      "grad_norm": 0.11463525891304016,
      "learning_rate": 2e-07,
      "loss": 0.0223,
      "step": 906
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08465312170052908,
      "grad_norm": 0.12273424863815308,
      "learning_rate": 2e-07,
      "loss": 0.0309,
      "step": 907
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08474645480052967,
      "grad_norm": 0.12761981785297394,
      "learning_rate": 2e-07,
      "loss": 0.0209,
      "step": 908
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08483978790053025,
      "grad_norm": 0.12849204242229462,
      "learning_rate": 2e-07,
      "loss": 0.0329,
      "step": 909
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08493312100053083,
      "grad_norm": 0.13091443479061127,
      "learning_rate": 2e-07,
      "loss": -0.0119,
      "step": 910
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08502645410053142,
      "grad_norm": 0.1361619532108307,
      "learning_rate": 2e-07,
      "loss": 0.0113,
      "step": 911
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.085119787200532,
      "grad_norm": 0.11390850692987442,
      "learning_rate": 2e-07,
      "loss": 0.0746,
      "step": 912
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08521312030053259,
      "grad_norm": 0.11825183033943176,
      "learning_rate": 2e-07,
      "loss": 0.0625,
      "step": 913
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08530645340053317,
      "grad_norm": 0.1295870989561081,
      "learning_rate": 2e-07,
      "loss": 0.0284,
      "step": 914
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08539978650053374,
      "grad_norm": 0.11985878646373749,
      "learning_rate": 2e-07,
      "loss": 0.0665,
      "step": 915
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08549311960053434,
      "grad_norm": 0.118943952023983,
      "learning_rate": 2e-07,
      "loss": 0.0489,
      "step": 916
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08558645270053492,
      "grad_norm": 0.1350439190864563,
      "learning_rate": 2e-07,
      "loss": 0.0474,
      "step": 917
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0856797858005355,
      "grad_norm": 0.11932313442230225,
      "learning_rate": 2e-07,
      "loss": 0.0262,
      "step": 918
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08577311890053609,
      "grad_norm": 0.11024530231952667,
      "learning_rate": 2e-07,
      "loss": 0.0309,
      "step": 919
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08586645200053666,
      "grad_norm": 0.11782639473676682,
      "learning_rate": 2e-07,
      "loss": 0.0105,
      "step": 920
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08595978510053724,
      "grad_norm": 0.12905670702457428,
      "learning_rate": 2e-07,
      "loss": -0.0011,
      "step": 921
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08605311820053783,
      "grad_norm": 0.1251768171787262,
      "learning_rate": 2e-07,
      "loss": 0.0531,
      "step": 922
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08614645130053841,
      "grad_norm": 0.11842817068099976,
      "learning_rate": 2e-07,
      "loss": 0.0083,
      "step": 923
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.086239784400539,
      "grad_norm": 0.12237225472927094,
      "learning_rate": 2e-07,
      "loss": 0.0174,
      "step": 924
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08633311750053958,
      "grad_norm": 0.11277849227190018,
      "learning_rate": 2e-07,
      "loss": 0.0261,
      "step": 925
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08642645060054016,
      "grad_norm": 0.11197086423635483,
      "learning_rate": 2e-07,
      "loss": 0.0356,
      "step": 926
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08651978370054075,
      "grad_norm": 0.19528403878211975,
      "learning_rate": 2e-07,
      "loss": 0.0454,
      "step": 927
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08661311680054133,
      "grad_norm": 0.12729953229427338,
      "learning_rate": 2e-07,
      "loss": 0.0159,
      "step": 928
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08670644990054191,
      "grad_norm": 0.12793217599391937,
      "learning_rate": 2e-07,
      "loss": 0.046,
      "step": 929
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0867997830005425,
      "grad_norm": 0.11462729424238205,
      "learning_rate": 2e-07,
      "loss": 0.003,
      "step": 930
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08689311610054308,
      "grad_norm": 0.11786491423845291,
      "learning_rate": 2e-07,
      "loss": 0.0147,
      "step": 931
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08698644920054366,
      "grad_norm": 0.12702345848083496,
      "learning_rate": 2e-07,
      "loss": 0.0343,
      "step": 932
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08707978230054425,
      "grad_norm": 0.1180480569601059,
      "learning_rate": 2e-07,
      "loss": 0.0417,
      "step": 933
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08717311540054483,
      "grad_norm": 0.11706763505935669,
      "learning_rate": 2e-07,
      "loss": 0.0586,
      "step": 934
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08726644850054542,
      "grad_norm": 0.14973102509975433,
      "learning_rate": 2e-07,
      "loss": -0.0179,
      "step": 935
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.087359781600546,
      "grad_norm": 0.12751786410808563,
      "learning_rate": 2e-07,
      "loss": -0.0014,
      "step": 936
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08745311470054658,
      "grad_norm": 0.12770016491413116,
      "learning_rate": 2e-07,
      "loss": 0.0779,
      "step": 937
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08754644780054717,
      "grad_norm": 0.11583378911018372,
      "learning_rate": 2e-07,
      "loss": 0.0166,
      "step": 938
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08763978090054775,
      "grad_norm": 0.1234976127743721,
      "learning_rate": 2e-07,
      "loss": 0.0623,
      "step": 939
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08773311400054833,
      "grad_norm": 0.12373754382133484,
      "learning_rate": 2e-07,
      "loss": 0.0113,
      "step": 940
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08782644710054892,
      "grad_norm": 0.12702524662017822,
      "learning_rate": 2e-07,
      "loss": 0.0505,
      "step": 941
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0879197802005495,
      "grad_norm": 0.12134741246700287,
      "learning_rate": 2e-07,
      "loss": 0.027,
      "step": 942
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08801311330055009,
      "grad_norm": 0.1295042335987091,
      "learning_rate": 2e-07,
      "loss": 0.0766,
      "step": 943
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08810644640055067,
      "grad_norm": 0.11991428583860397,
      "learning_rate": 2e-07,
      "loss": 0.0078,
      "step": 944
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08819977950055125,
      "grad_norm": 0.12392047792673111,
      "learning_rate": 2e-07,
      "loss": 0.0931,
      "step": 945
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08829311260055184,
      "grad_norm": 0.13601675629615784,
      "learning_rate": 2e-07,
      "loss": 0.017,
      "step": 946
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08838644570055242,
      "grad_norm": 0.12515506148338318,
      "learning_rate": 2e-07,
      "loss": 0.0152,
      "step": 947
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.088479778800553,
      "grad_norm": 0.14124850928783417,
      "learning_rate": 2e-07,
      "loss": 0.0055,
      "step": 948
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08857311190055359,
      "grad_norm": 0.13649940490722656,
      "learning_rate": 2e-07,
      "loss": 0.0017,
      "step": 949
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08866644500055416,
      "grad_norm": 0.12156989425420761,
      "learning_rate": 2e-07,
      "loss": -0.0188,
      "step": 950
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08875977810055474,
      "grad_norm": 0.1219983845949173,
      "learning_rate": 2e-07,
      "loss": 0.0455,
      "step": 951
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08885311120055533,
      "grad_norm": 0.13010255992412567,
      "learning_rate": 2e-07,
      "loss": -0.001,
      "step": 952
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08894644430055591,
      "grad_norm": 0.1486521065235138,
      "learning_rate": 2e-07,
      "loss": 0.0179,
      "step": 953
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0890397774005565,
      "grad_norm": 0.12371961772441864,
      "learning_rate": 2e-07,
      "loss": 0.0225,
      "step": 954
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08913311050055708,
      "grad_norm": 0.10351689904928207,
      "learning_rate": 2e-07,
      "loss": 0.0148,
      "step": 955
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08922644360055766,
      "grad_norm": 0.14067280292510986,
      "learning_rate": 2e-07,
      "loss": 0.0402,
      "step": 956
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08931977670055825,
      "grad_norm": 0.12496519088745117,
      "learning_rate": 2e-07,
      "loss": -0.0029,
      "step": 957
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08941310980055883,
      "grad_norm": 0.12028787285089493,
      "learning_rate": 2e-07,
      "loss": 0.0401,
      "step": 958
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08950644290055941,
      "grad_norm": 0.11538785696029663,
      "learning_rate": 2e-07,
      "loss": 0.0085,
      "step": 959
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08959977600056,
      "grad_norm": 0.121493399143219,
      "learning_rate": 2e-07,
      "loss": 0.0532,
      "step": 960
    },
    {
      "clip_ratio/high_max": 0.010437682183692232,
      "clip_ratio/high_mean": 0.0051509564218577,
      "clip_ratio/low_mean": 0.0035535662100301124,
      "clip_ratio/low_min": 0.00044791259279008955,
      "clip_ratio/region_mean": 0.008704522508196533,
      "completions/clipped_ratio": 0.013218470982142905,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4049.0,
      "completions/mean_length": 613.3426513671875,
      "completions/mean_terminated_length": 566.6906127929688,
      "completions/min_length": 77.0,
      "completions/min_terminated_length": 77.0,
      "epoch": 0.08969310910056058,
      "grad_norm": 0.11790748685598373,
      "learning_rate": 2e-07,
      "loss": 0.0131,
      "num_tokens": 654088676.0,
      "reward": 0.5857980251312256,
      "reward_std": 0.2002045214176178,
      "rewards/simpleverify_reward/mean": 0.5857979655265808,
      "rewards/simpleverify_reward/std": 0.49258795380592346,
      "step": 961
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08978644220056116,
      "grad_norm": 0.12580302357673645,
      "learning_rate": 2e-07,
      "loss": 0.0554,
      "step": 962
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08987977530056175,
      "grad_norm": 0.12604328989982605,
      "learning_rate": 2e-07,
      "loss": 0.0033,
      "step": 963
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.08997310840056233,
      "grad_norm": 0.12717796862125397,
      "learning_rate": 2e-07,
      "loss": -0.0242,
      "step": 964
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09006644150056292,
      "grad_norm": 0.13788248598575592,
      "learning_rate": 2e-07,
      "loss": 0.0084,
      "step": 965
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0901597746005635,
      "grad_norm": 0.11159532517194748,
      "learning_rate": 2e-07,
      "loss": 0.0279,
      "step": 966
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09025310770056408,
      "grad_norm": 0.13069631159305573,
      "learning_rate": 2e-07,
      "loss": 0.0046,
      "step": 967
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09034644080056467,
      "grad_norm": 0.12297243624925613,
      "learning_rate": 2e-07,
      "loss": 0.0487,
      "step": 968
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09043977390056525,
      "grad_norm": 0.13381244242191315,
      "learning_rate": 2e-07,
      "loss": 0.0543,
      "step": 969
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09053310700056583,
      "grad_norm": 0.1314571052789688,
      "learning_rate": 2e-07,
      "loss": -0.0218,
      "step": 970
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09062644010056642,
      "grad_norm": 0.1321249008178711,
      "learning_rate": 2e-07,
      "loss": 0.0477,
      "step": 971
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.090719773200567,
      "grad_norm": 0.12110405415296555,
      "learning_rate": 2e-07,
      "loss": 0.0448,
      "step": 972
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09081310630056758,
      "grad_norm": 0.11183035373687744,
      "learning_rate": 2e-07,
      "loss": 0.0389,
      "step": 973
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09090643940056817,
      "grad_norm": 0.1100270226597786,
      "learning_rate": 2e-07,
      "loss": 0.0172,
      "step": 974
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09099977250056875,
      "grad_norm": 0.14608921110630035,
      "learning_rate": 2e-07,
      "loss": 0.042,
      "step": 975
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09109310560056934,
      "grad_norm": 0.24353210628032684,
      "learning_rate": 2e-07,
      "loss": 0.0256,
      "step": 976
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09118643870056992,
      "grad_norm": 0.13845276832580566,
      "learning_rate": 2e-07,
      "loss": 0.0278,
      "step": 977
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0912797718005705,
      "grad_norm": 0.10552609711885452,
      "learning_rate": 2e-07,
      "loss": 0.0217,
      "step": 978
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09137310490057109,
      "grad_norm": 0.13302023708820343,
      "learning_rate": 2e-07,
      "loss": 0.0148,
      "step": 979
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09146643800057166,
      "grad_norm": 0.1335231512784958,
      "learning_rate": 2e-07,
      "loss": -0.0143,
      "step": 980
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09155977110057224,
      "grad_norm": 0.11808809638023376,
      "learning_rate": 2e-07,
      "loss": 0.0531,
      "step": 981
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09165310420057284,
      "grad_norm": 0.12824606895446777,
      "learning_rate": 2e-07,
      "loss": 0.0357,
      "step": 982
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09174643730057341,
      "grad_norm": 0.1350392997264862,
      "learning_rate": 2e-07,
      "loss": 0.0734,
      "step": 983
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09183977040057399,
      "grad_norm": 0.12262378633022308,
      "learning_rate": 2e-07,
      "loss": 0.0402,
      "step": 984
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09193310350057458,
      "grad_norm": 0.12261346727609634,
      "learning_rate": 2e-07,
      "loss": 0.0163,
      "step": 985
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09202643660057516,
      "grad_norm": 0.12925438582897186,
      "learning_rate": 2e-07,
      "loss": 0.046,
      "step": 986
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09211976970057575,
      "grad_norm": 0.1323062628507614,
      "learning_rate": 2e-07,
      "loss": 0.0801,
      "step": 987
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09221310280057633,
      "grad_norm": 0.12085221707820892,
      "learning_rate": 2e-07,
      "loss": 0.0439,
      "step": 988
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09230643590057691,
      "grad_norm": 0.12385974824428558,
      "learning_rate": 2e-07,
      "loss": 0.0125,
      "step": 989
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0923997690005775,
      "grad_norm": 0.14545592665672302,
      "learning_rate": 2e-07,
      "loss": -0.0456,
      "step": 990
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09249310210057808,
      "grad_norm": 0.1537473350763321,
      "learning_rate": 2e-07,
      "loss": 0.0719,
      "step": 991
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09258643520057866,
      "grad_norm": 0.13486535847187042,
      "learning_rate": 2e-07,
      "loss": -0.0114,
      "step": 992
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09267976830057925,
      "grad_norm": 0.1258154660463333,
      "learning_rate": 2e-07,
      "loss": 0.0479,
      "step": 993
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09277310140057983,
      "grad_norm": 0.13114792108535767,
      "learning_rate": 2e-07,
      "loss": 0.046,
      "step": 994
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09286643450058042,
      "grad_norm": 0.13173554837703705,
      "learning_rate": 2e-07,
      "loss": 0.0251,
      "step": 995
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.092959767600581,
      "grad_norm": 0.13012327253818512,
      "learning_rate": 2e-07,
      "loss": -0.0174,
      "step": 996
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09305310070058158,
      "grad_norm": 0.12421735376119614,
      "learning_rate": 2e-07,
      "loss": 0.0358,
      "step": 997
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09314643380058217,
      "grad_norm": 0.12075912952423096,
      "learning_rate": 2e-07,
      "loss": 0.0116,
      "step": 998
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09323976690058275,
      "grad_norm": 0.12936711311340332,
      "learning_rate": 2e-07,
      "loss": 0.0216,
      "step": 999
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09333310000058333,
      "grad_norm": 0.131841242313385,
      "learning_rate": 2e-07,
      "loss": 0.0509,
      "step": 1000
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09342643310058392,
      "grad_norm": 0.12027263641357422,
      "learning_rate": 2e-07,
      "loss": 0.0358,
      "step": 1001
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0935197662005845,
      "grad_norm": 0.13181601464748383,
      "learning_rate": 2e-07,
      "loss": -0.0164,
      "step": 1002
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09361309930058508,
      "grad_norm": 0.11895671486854553,
      "learning_rate": 2e-07,
      "loss": 0.0009,
      "step": 1003
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09370643240058567,
      "grad_norm": 0.12323889881372452,
      "learning_rate": 2e-07,
      "loss": 0.0323,
      "step": 1004
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09379976550058625,
      "grad_norm": 0.13781878352165222,
      "learning_rate": 2e-07,
      "loss": 0.0301,
      "step": 1005
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09389309860058684,
      "grad_norm": 0.12690988183021545,
      "learning_rate": 2e-07,
      "loss": 0.0072,
      "step": 1006
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09398643170058742,
      "grad_norm": 0.14287704229354858,
      "learning_rate": 2e-07,
      "loss": 0.0427,
      "step": 1007
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.094079764800588,
      "grad_norm": 0.12391387671232224,
      "learning_rate": 2e-07,
      "loss": 0.0214,
      "step": 1008
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09417309790058859,
      "grad_norm": 0.1258545070886612,
      "learning_rate": 2e-07,
      "loss": 0.0312,
      "step": 1009
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09426643100058917,
      "grad_norm": 0.13660462200641632,
      "learning_rate": 2e-07,
      "loss": 0.0565,
      "step": 1010
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09435976410058974,
      "grad_norm": 0.1260143518447876,
      "learning_rate": 2e-07,
      "loss": 0.0246,
      "step": 1011
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09445309720059034,
      "grad_norm": 0.12764915823936462,
      "learning_rate": 2e-07,
      "loss": 0.0397,
      "step": 1012
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09454643030059091,
      "grad_norm": 0.1330089122056961,
      "learning_rate": 2e-07,
      "loss": 0.025,
      "step": 1013
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09463976340059149,
      "grad_norm": 0.12474074959754944,
      "learning_rate": 2e-07,
      "loss": 0.0636,
      "step": 1014
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09473309650059208,
      "grad_norm": 0.1148066520690918,
      "learning_rate": 2e-07,
      "loss": 0.0544,
      "step": 1015
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09482642960059266,
      "grad_norm": 0.1243598684668541,
      "learning_rate": 2e-07,
      "loss": 0.0361,
      "step": 1016
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09491976270059325,
      "grad_norm": 0.12512850761413574,
      "learning_rate": 2e-07,
      "loss": -0.001,
      "step": 1017
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09501309580059383,
      "grad_norm": 0.12310733646154404,
      "learning_rate": 2e-07,
      "loss": 0.0651,
      "step": 1018
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09510642890059441,
      "grad_norm": 0.12769833207130432,
      "learning_rate": 2e-07,
      "loss": 0.037,
      "step": 1019
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.095199762000595,
      "grad_norm": 0.13213273882865906,
      "learning_rate": 2e-07,
      "loss": 0.0175,
      "step": 1020
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09529309510059558,
      "grad_norm": 0.11583854258060455,
      "learning_rate": 2e-07,
      "loss": 0.0038,
      "step": 1021
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09538642820059616,
      "grad_norm": 0.1235031858086586,
      "learning_rate": 2e-07,
      "loss": 0.0518,
      "step": 1022
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09547976130059675,
      "grad_norm": 0.12434105575084686,
      "learning_rate": 2e-07,
      "loss": 0.0566,
      "step": 1023
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09557309440059733,
      "grad_norm": 0.11817721277475357,
      "learning_rate": 2e-07,
      "loss": 0.0169,
      "step": 1024
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.014177594866071397,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4095.0,
      "completions/mean_length": 622.6942138671875,
      "completions/mean_terminated_length": 572.742919921875,
      "completions/min_length": 83.0,
      "completions/min_terminated_length": 83.0,
      "epoch": 0.09566642750059791,
      "grad_norm": 0.1316581666469574,
      "learning_rate": 2e-07,
      "loss": -0.0141,
      "num_tokens": 695637763.0,
      "reward": 0.5737479329109192,
      "reward_std": 0.19330938160419464,
      "rewards/simpleverify_reward/mean": 0.5737479329109192,
      "rewards/simpleverify_reward/std": 0.4945356249809265,
      "step": 1025
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0957597606005985,
      "grad_norm": 0.12498697638511658,
      "learning_rate": 2e-07,
      "loss": 0.0301,
      "step": 1026
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09585309370059908,
      "grad_norm": 0.13646142184734344,
      "learning_rate": 2e-07,
      "loss": 0.0424,
      "step": 1027
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09594642680059967,
      "grad_norm": 0.12600235641002655,
      "learning_rate": 2e-07,
      "loss": -0.058,
      "step": 1028
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09603975990060025,
      "grad_norm": 0.12201765924692154,
      "learning_rate": 2e-07,
      "loss": 0.0038,
      "step": 1029
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09613309300060083,
      "grad_norm": 0.11507520079612732,
      "learning_rate": 2e-07,
      "loss": -0.0047,
      "step": 1030
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09622642610060142,
      "grad_norm": 0.12242674827575684,
      "learning_rate": 2e-07,
      "loss": 0.0513,
      "step": 1031
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.096319759200602,
      "grad_norm": 0.13105055689811707,
      "learning_rate": 2e-07,
      "loss": 0.0651,
      "step": 1032
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09641309230060258,
      "grad_norm": 0.11185221374034882,
      "learning_rate": 2e-07,
      "loss": 0.0484,
      "step": 1033
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09650642540060317,
      "grad_norm": 0.13074176013469696,
      "learning_rate": 2e-07,
      "loss": 0.022,
      "step": 1034
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09659975850060375,
      "grad_norm": 0.10907220095396042,
      "learning_rate": 2e-07,
      "loss": 0.0097,
      "step": 1035
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09669309160060433,
      "grad_norm": 0.11873073875904083,
      "learning_rate": 2e-07,
      "loss": 0.0114,
      "step": 1036
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09678642470060492,
      "grad_norm": 0.11945661157369614,
      "learning_rate": 2e-07,
      "loss": -0.004,
      "step": 1037
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0968797578006055,
      "grad_norm": 0.1256019026041031,
      "learning_rate": 2e-07,
      "loss": 0.0505,
      "step": 1038
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09697309090060609,
      "grad_norm": 0.13739430904388428,
      "learning_rate": 2e-07,
      "loss": 0.0016,
      "step": 1039
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09706642400060667,
      "grad_norm": 0.12666425108909607,
      "learning_rate": 2e-07,
      "loss": 0.0507,
      "step": 1040
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09715975710060724,
      "grad_norm": 0.11463279277086258,
      "learning_rate": 2e-07,
      "loss": 0.0116,
      "step": 1041
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09725309020060784,
      "grad_norm": 0.12110929936170578,
      "learning_rate": 2e-07,
      "loss": 0.0401,
      "step": 1042
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09734642330060841,
      "grad_norm": 0.1359110325574875,
      "learning_rate": 2e-07,
      "loss": 0.0182,
      "step": 1043
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09743975640060899,
      "grad_norm": 0.11389145255088806,
      "learning_rate": 2e-07,
      "loss": 0.0293,
      "step": 1044
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09753308950060958,
      "grad_norm": 0.1231398656964302,
      "learning_rate": 2e-07,
      "loss": -0.0253,
      "step": 1045
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09762642260061016,
      "grad_norm": 0.17772337794303894,
      "learning_rate": 2e-07,
      "loss": 0.0316,
      "step": 1046
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09771975570061076,
      "grad_norm": 0.11407757550477982,
      "learning_rate": 2e-07,
      "loss": 0.0395,
      "step": 1047
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09781308880061133,
      "grad_norm": 0.1250477433204651,
      "learning_rate": 2e-07,
      "loss": 0.0315,
      "step": 1048
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09790642190061191,
      "grad_norm": 0.1850556880235672,
      "learning_rate": 2e-07,
      "loss": 0.0119,
      "step": 1049
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0979997550006125,
      "grad_norm": 0.1270701289176941,
      "learning_rate": 2e-07,
      "loss": -0.0083,
      "step": 1050
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09809308810061308,
      "grad_norm": 0.1168687492609024,
      "learning_rate": 2e-07,
      "loss": 0.0472,
      "step": 1051
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09818642120061366,
      "grad_norm": 0.12744373083114624,
      "learning_rate": 2e-07,
      "loss": 0.0008,
      "step": 1052
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09827975430061425,
      "grad_norm": 0.1210247203707695,
      "learning_rate": 2e-07,
      "loss": 0.0268,
      "step": 1053
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09837308740061483,
      "grad_norm": 0.12781542539596558,
      "learning_rate": 2e-07,
      "loss": 0.0288,
      "step": 1054
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09846642050061541,
      "grad_norm": 0.14108054339885712,
      "learning_rate": 2e-07,
      "loss": 0.0538,
      "step": 1055
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.098559753600616,
      "grad_norm": 0.11946498602628708,
      "learning_rate": 2e-07,
      "loss": 0.025,
      "step": 1056
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09865308670061658,
      "grad_norm": 0.33612948656082153,
      "learning_rate": 2e-07,
      "loss": 0.0307,
      "step": 1057
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09874641980061717,
      "grad_norm": 0.1263088434934616,
      "learning_rate": 2e-07,
      "loss": 0.014,
      "step": 1058
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09883975290061775,
      "grad_norm": 0.1360570341348648,
      "learning_rate": 2e-07,
      "loss": 0.0537,
      "step": 1059
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09893308600061833,
      "grad_norm": 0.11681803315877914,
      "learning_rate": 2e-07,
      "loss": 0.0073,
      "step": 1060
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09902641910061892,
      "grad_norm": 0.12059196829795837,
      "learning_rate": 2e-07,
      "loss": 0.0361,
      "step": 1061
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.0991197522006195,
      "grad_norm": 0.13508820533752441,
      "learning_rate": 2e-07,
      "loss": 0.0639,
      "step": 1062
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09921308530062008,
      "grad_norm": 0.12632830440998077,
      "learning_rate": 2e-07,
      "loss": 0.0638,
      "step": 1063
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09930641840062067,
      "grad_norm": 0.12242299318313599,
      "learning_rate": 2e-07,
      "loss": 0.0208,
      "step": 1064
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09939975150062125,
      "grad_norm": 0.1329362392425537,
      "learning_rate": 2e-07,
      "loss": 0.0094,
      "step": 1065
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09949308460062183,
      "grad_norm": 0.12316466867923737,
      "learning_rate": 2e-07,
      "loss": 0.0322,
      "step": 1066
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09958641770062242,
      "grad_norm": 0.11811467260122299,
      "learning_rate": 2e-07,
      "loss": -0.0181,
      "step": 1067
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.099679750800623,
      "grad_norm": 0.1136048287153244,
      "learning_rate": 2e-07,
      "loss": 0.0292,
      "step": 1068
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09977308390062359,
      "grad_norm": 0.14453837275505066,
      "learning_rate": 2e-07,
      "loss": -0.0128,
      "step": 1069
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09986641700062417,
      "grad_norm": 0.1360822319984436,
      "learning_rate": 2e-07,
      "loss": 0.0367,
      "step": 1070
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.09995975010062474,
      "grad_norm": 0.12935367226600647,
      "learning_rate": 2e-07,
      "loss": 0.0279,
      "step": 1071
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10005308320062534,
      "grad_norm": 0.12355770915746689,
      "learning_rate": 2e-07,
      "loss": 0.0386,
      "step": 1072
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10014641630062592,
      "grad_norm": 0.1192934438586235,
      "learning_rate": 2e-07,
      "loss": 0.0574,
      "step": 1073
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1002397494006265,
      "grad_norm": 0.129846453666687,
      "learning_rate": 2e-07,
      "loss": 0.0484,
      "step": 1074
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10033308250062709,
      "grad_norm": 0.1394384503364563,
      "learning_rate": 2e-07,
      "loss": 0.0356,
      "step": 1075
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10042641560062766,
      "grad_norm": 0.12558715045452118,
      "learning_rate": 2e-07,
      "loss": 0.0689,
      "step": 1076
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10051974870062824,
      "grad_norm": 0.1378021091222763,
      "learning_rate": 2e-07,
      "loss": 0.0504,
      "step": 1077
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10061308180062883,
      "grad_norm": 0.11771351844072342,
      "learning_rate": 2e-07,
      "loss": 0.0605,
      "step": 1078
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10070641490062941,
      "grad_norm": 0.124101422727108,
      "learning_rate": 2e-07,
      "loss": 0.069,
      "step": 1079
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10079974800063,
      "grad_norm": 0.12010723352432251,
      "learning_rate": 2e-07,
      "loss": -0.0035,
      "step": 1080
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10089308110063058,
      "grad_norm": 0.1290493756532669,
      "learning_rate": 2e-07,
      "loss": 0.0495,
      "step": 1081
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10098641420063116,
      "grad_norm": 0.11392639577388763,
      "learning_rate": 2e-07,
      "loss": 0.0006,
      "step": 1082
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10107974730063175,
      "grad_norm": 0.12000121176242828,
      "learning_rate": 2e-07,
      "loss": 0.018,
      "step": 1083
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10117308040063233,
      "grad_norm": 0.12304224818944931,
      "learning_rate": 2e-07,
      "loss": 0.0629,
      "step": 1084
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10126641350063291,
      "grad_norm": 0.16488148272037506,
      "learning_rate": 2e-07,
      "loss": 0.0297,
      "step": 1085
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1013597466006335,
      "grad_norm": 0.1308477520942688,
      "learning_rate": 2e-07,
      "loss": 0.0248,
      "step": 1086
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10145307970063408,
      "grad_norm": 0.12387493252754211,
      "learning_rate": 2e-07,
      "loss": -0.0092,
      "step": 1087
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10154641280063466,
      "grad_norm": 0.11677070707082748,
      "learning_rate": 2e-07,
      "loss": 0.0413,
      "step": 1088
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0135498046875,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4075.0,
      "completions/mean_length": 618.5260620117188,
      "completions/mean_terminated_length": 570.7597045898438,
      "completions/min_length": 91.0,
      "completions/min_terminated_length": 91.0,
      "epoch": 0.10163974590063525,
      "grad_norm": 0.129710853099823,
      "learning_rate": 2e-07,
      "loss": 0.0266,
      "num_tokens": 736832311.0,
      "reward": 0.5871233344078064,
      "reward_std": 0.19259746372699738,
      "rewards/simpleverify_reward/mean": 0.5871233344078064,
      "rewards/simpleverify_reward/std": 0.4923553168773651,
      "step": 1089
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10173307900063583,
      "grad_norm": 0.12425310909748077,
      "learning_rate": 2e-07,
      "loss": 0.0202,
      "step": 1090
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10182641210063642,
      "grad_norm": 0.11199373751878738,
      "learning_rate": 2e-07,
      "loss": 0.062,
      "step": 1091
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.101919745200637,
      "grad_norm": 0.1283552348613739,
      "learning_rate": 2e-07,
      "loss": 0.0056,
      "step": 1092
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10201307830063758,
      "grad_norm": 0.12982302904129028,
      "learning_rate": 2e-07,
      "loss": 0.0258,
      "step": 1093
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10210641140063817,
      "grad_norm": 0.1246814951300621,
      "learning_rate": 2e-07,
      "loss": 0.0061,
      "step": 1094
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10219974450063875,
      "grad_norm": 0.12260641157627106,
      "learning_rate": 2e-07,
      "loss": 0.0315,
      "step": 1095
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10229307760063933,
      "grad_norm": 0.12535782158374786,
      "learning_rate": 2e-07,
      "loss": 0.0385,
      "step": 1096
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10238641070063992,
      "grad_norm": 0.11780212819576263,
      "learning_rate": 2e-07,
      "loss": 0.0473,
      "step": 1097
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1024797438006405,
      "grad_norm": 0.12184536457061768,
      "learning_rate": 2e-07,
      "loss": 0.0091,
      "step": 1098
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10257307690064109,
      "grad_norm": 0.16125360131263733,
      "learning_rate": 2e-07,
      "loss": 0.0209,
      "step": 1099
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10266641000064167,
      "grad_norm": 0.129767045378685,
      "learning_rate": 2e-07,
      "loss": 0.0204,
      "step": 1100
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10275974310064225,
      "grad_norm": 0.12149225920438766,
      "learning_rate": 2e-07,
      "loss": 0.0449,
      "step": 1101
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10285307620064284,
      "grad_norm": 0.11585115641355515,
      "learning_rate": 2e-07,
      "loss": 0.0016,
      "step": 1102
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10294640930064342,
      "grad_norm": 0.11527204513549805,
      "learning_rate": 2e-07,
      "loss": 0.0237,
      "step": 1103
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.103039742400644,
      "grad_norm": 0.12590822577476501,
      "learning_rate": 2e-07,
      "loss": 0.0478,
      "step": 1104
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10313307550064459,
      "grad_norm": 0.13409964740276337,
      "learning_rate": 2e-07,
      "loss": 0.0078,
      "step": 1105
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10322640860064516,
      "grad_norm": 0.1281125694513321,
      "learning_rate": 2e-07,
      "loss": 0.0423,
      "step": 1106
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10331974170064574,
      "grad_norm": 0.13257557153701782,
      "learning_rate": 2e-07,
      "loss": 0.005,
      "step": 1107
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10341307480064633,
      "grad_norm": 0.11419656872749329,
      "learning_rate": 2e-07,
      "loss": 0.0004,
      "step": 1108
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10350640790064691,
      "grad_norm": 0.12049297243356705,
      "learning_rate": 2e-07,
      "loss": 0.0565,
      "step": 1109
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1035997410006475,
      "grad_norm": 0.15598013997077942,
      "learning_rate": 2e-07,
      "loss": 0.0524,
      "step": 1110
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10369307410064808,
      "grad_norm": 0.13944809138774872,
      "learning_rate": 2e-07,
      "loss": 0.024,
      "step": 1111
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10378640720064866,
      "grad_norm": 0.12656055390834808,
      "learning_rate": 2e-07,
      "loss": 0.0231,
      "step": 1112
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10387974030064925,
      "grad_norm": 0.14490559697151184,
      "learning_rate": 2e-07,
      "loss": 0.0587,
      "step": 1113
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10397307340064983,
      "grad_norm": 0.1195591539144516,
      "learning_rate": 2e-07,
      "loss": 0.0106,
      "step": 1114
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10406640650065041,
      "grad_norm": 0.1294788271188736,
      "learning_rate": 2e-07,
      "loss": 0.0526,
      "step": 1115
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.104159739600651,
      "grad_norm": 0.13575363159179688,
      "learning_rate": 2e-07,
      "loss": 0.0186,
      "step": 1116
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10425307270065158,
      "grad_norm": 0.1433030515909195,
      "learning_rate": 2e-07,
      "loss": 0.0229,
      "step": 1117
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10434640580065216,
      "grad_norm": 0.1215771734714508,
      "learning_rate": 2e-07,
      "loss": -0.0045,
      "step": 1118
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10443973890065275,
      "grad_norm": 0.12646552920341492,
      "learning_rate": 2e-07,
      "loss": 0.0145,
      "step": 1119
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10453307200065333,
      "grad_norm": 0.1285991221666336,
      "learning_rate": 2e-07,
      "loss": 0.0268,
      "step": 1120
    },
    {
      "clip_ratio/high_max": 0.011275653843767941,
      "clip_ratio/high_mean": 0.00543773511890322,
      "clip_ratio/low_mean": 0.003624753844633233,
      "clip_ratio/low_min": 0.0002658166085893754,
      "clip_ratio/region_mean": 0.00906248899991624,
      "completions/clipped_ratio": 0.013636997767857095,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4083.0,
      "completions/mean_length": 621.5830078125,
      "completions/mean_terminated_length": 573.54736328125,
      "completions/min_length": 5.0,
      "completions/min_terminated_length": 5.0,
      "epoch": 0.10462640510065392,
      "grad_norm": 0.12258480489253998,
      "learning_rate": 2e-07,
      "loss": 0.0079,
      "num_tokens": 778202160.0,
      "reward": 0.5866001844406128,
      "reward_std": 0.19235768914222717,
      "rewards/simpleverify_reward/mean": 0.5866001844406128,
      "rewards/simpleverify_reward/std": 0.4924476146697998,
      "step": 1121
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1047197382006545,
      "grad_norm": 0.12275806814432144,
      "learning_rate": 2e-07,
      "loss": 0.0424,
      "step": 1122
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10481307130065508,
      "grad_norm": 0.1303173154592514,
      "learning_rate": 2e-07,
      "loss": 0.0309,
      "step": 1123
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10490640440065567,
      "grad_norm": 0.12321112304925919,
      "learning_rate": 2e-07,
      "loss": -0.0192,
      "step": 1124
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10499973750065625,
      "grad_norm": 0.1279294639825821,
      "learning_rate": 2e-07,
      "loss": -0.0149,
      "step": 1125
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10509307060065683,
      "grad_norm": 0.12685002386569977,
      "learning_rate": 2e-07,
      "loss": 0.0356,
      "step": 1126
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10518640370065742,
      "grad_norm": 0.13565027713775635,
      "learning_rate": 2e-07,
      "loss": -0.0219,
      "step": 1127
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.105279736800658,
      "grad_norm": 0.11720327287912369,
      "learning_rate": 2e-07,
      "loss": 0.028,
      "step": 1128
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10537306990065858,
      "grad_norm": 0.1343325823545456,
      "learning_rate": 2e-07,
      "loss": 0.0607,
      "step": 1129
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10546640300065917,
      "grad_norm": 0.12469281256198883,
      "learning_rate": 2e-07,
      "loss": -0.0125,
      "step": 1130
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10555973610065975,
      "grad_norm": 0.12600409984588623,
      "learning_rate": 2e-07,
      "loss": -0.0212,
      "step": 1131
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10565306920066034,
      "grad_norm": 0.12666453421115875,
      "learning_rate": 2e-07,
      "loss": 0.042,
      "step": 1132
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10574640230066092,
      "grad_norm": 0.12581267952919006,
      "learning_rate": 2e-07,
      "loss": 0.0004,
      "step": 1133
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1058397354006615,
      "grad_norm": 0.11753392219543457,
      "learning_rate": 2e-07,
      "loss": 0.0252,
      "step": 1134
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10593306850066209,
      "grad_norm": 0.12723000347614288,
      "learning_rate": 2e-07,
      "loss": 0.0637,
      "step": 1135
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10602640160066266,
      "grad_norm": 0.12447874993085861,
      "learning_rate": 2e-07,
      "loss": 0.0335,
      "step": 1136
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10611973470066324,
      "grad_norm": 0.14147357642650604,
      "learning_rate": 2e-07,
      "loss": 0.0608,
      "step": 1137
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10621306780066384,
      "grad_norm": 0.13559502363204956,
      "learning_rate": 2e-07,
      "loss": 0.0269,
      "step": 1138
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10630640090066441,
      "grad_norm": 0.10702571272850037,
      "learning_rate": 2e-07,
      "loss": 0.0123,
      "step": 1139
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10639973400066499,
      "grad_norm": 0.12850673496723175,
      "learning_rate": 2e-07,
      "loss": 0.0162,
      "step": 1140
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10649306710066558,
      "grad_norm": 0.11446702480316162,
      "learning_rate": 2e-07,
      "loss": 0.005,
      "step": 1141
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10658640020066616,
      "grad_norm": 0.12650370597839355,
      "learning_rate": 2e-07,
      "loss": 0.002,
      "step": 1142
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10667973330066675,
      "grad_norm": 0.1344020813703537,
      "learning_rate": 2e-07,
      "loss": -0.0006,
      "step": 1143
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10677306640066733,
      "grad_norm": 0.12671393156051636,
      "learning_rate": 2e-07,
      "loss": 0.007,
      "step": 1144
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10686639950066791,
      "grad_norm": 0.12316898256540298,
      "learning_rate": 2e-07,
      "loss": 0.0167,
      "step": 1145
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1069597326006685,
      "grad_norm": 0.12120793759822845,
      "learning_rate": 2e-07,
      "loss": 0.0165,
      "step": 1146
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10705306570066908,
      "grad_norm": 0.12227485328912735,
      "learning_rate": 2e-07,
      "loss": 0.014,
      "step": 1147
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10714639880066966,
      "grad_norm": 0.1323586404323578,
      "learning_rate": 2e-07,
      "loss": 0.0436,
      "step": 1148
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10723973190067025,
      "grad_norm": 0.12394601851701736,
      "learning_rate": 2e-07,
      "loss": 0.0321,
      "step": 1149
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10733306500067083,
      "grad_norm": 0.12051504105329514,
      "learning_rate": 2e-07,
      "loss": 0.0496,
      "step": 1150
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10742639810067142,
      "grad_norm": 0.12830033898353577,
      "learning_rate": 2e-07,
      "loss": 0.0517,
      "step": 1151
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.107519731200672,
      "grad_norm": 0.1275833398103714,
      "learning_rate": 2e-07,
      "loss": -0.0025,
      "step": 1152
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10761306430067258,
      "grad_norm": 0.12625139951705933,
      "learning_rate": 2e-07,
      "loss": 0.0164,
      "step": 1153
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10770639740067317,
      "grad_norm": 0.12693054974079132,
      "learning_rate": 2e-07,
      "loss": 0.0518,
      "step": 1154
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10779973050067375,
      "grad_norm": 0.12615393102169037,
      "learning_rate": 2e-07,
      "loss": 0.0444,
      "step": 1155
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10789306360067433,
      "grad_norm": 0.13960233330726624,
      "learning_rate": 2e-07,
      "loss": -0.0057,
      "step": 1156
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10798639670067492,
      "grad_norm": 0.14542271196842194,
      "learning_rate": 2e-07,
      "loss": 0.0071,
      "step": 1157
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1080797298006755,
      "grad_norm": 0.1156541183590889,
      "learning_rate": 2e-07,
      "loss": 0.0337,
      "step": 1158
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10817306290067608,
      "grad_norm": 0.1293029934167862,
      "learning_rate": 2e-07,
      "loss": 0.0044,
      "step": 1159
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10826639600067667,
      "grad_norm": 0.13470028340816498,
      "learning_rate": 2e-07,
      "loss": 0.0357,
      "step": 1160
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10835972910067725,
      "grad_norm": 0.11810430884361267,
      "learning_rate": 2e-07,
      "loss": 0.0222,
      "step": 1161
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10845306220067784,
      "grad_norm": 0.12567394971847534,
      "learning_rate": 2e-07,
      "loss": 0.0182,
      "step": 1162
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10854639530067842,
      "grad_norm": 0.12532854080200195,
      "learning_rate": 2e-07,
      "loss": 0.0081,
      "step": 1163
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.108639728400679,
      "grad_norm": 0.12335190176963806,
      "learning_rate": 2e-07,
      "loss": 0.0085,
      "step": 1164
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10873306150067959,
      "grad_norm": 0.11937543749809265,
      "learning_rate": 2e-07,
      "loss": 0.028,
      "step": 1165
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10882639460068017,
      "grad_norm": 0.1282988339662552,
      "learning_rate": 2e-07,
      "loss": 0.0289,
      "step": 1166
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10891972770068074,
      "grad_norm": 0.11780912429094315,
      "learning_rate": 2e-07,
      "loss": 0.0319,
      "step": 1167
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10901306080068134,
      "grad_norm": 0.13476280868053436,
      "learning_rate": 2e-07,
      "loss": 0.0439,
      "step": 1168
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10910639390068191,
      "grad_norm": 0.1252293437719345,
      "learning_rate": 2e-07,
      "loss": 0.0191,
      "step": 1169
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10919972700068249,
      "grad_norm": 0.11742155253887177,
      "learning_rate": 2e-07,
      "loss": 0.0373,
      "step": 1170
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10929306010068308,
      "grad_norm": 0.11951114982366562,
      "learning_rate": 2e-07,
      "loss": 0.0836,
      "step": 1171
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10938639320068366,
      "grad_norm": 0.12027372419834137,
      "learning_rate": 2e-07,
      "loss": 0.0779,
      "step": 1172
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10947972630068425,
      "grad_norm": 0.12384378910064697,
      "learning_rate": 2e-07,
      "loss": 0.0222,
      "step": 1173
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10957305940068483,
      "grad_norm": 0.12592965364456177,
      "learning_rate": 2e-07,
      "loss": 0.037,
      "step": 1174
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10966639250068541,
      "grad_norm": 0.14081540703773499,
      "learning_rate": 2e-07,
      "loss": 0.1091,
      "step": 1175
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.109759725600686,
      "grad_norm": 0.12784026563167572,
      "learning_rate": 2e-07,
      "loss": 0.0305,
      "step": 1176
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10985305870068658,
      "grad_norm": 0.1324957311153412,
      "learning_rate": 2e-07,
      "loss": -0.0032,
      "step": 1177
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.10994639180068716,
      "grad_norm": 0.29785025119781494,
      "learning_rate": 2e-07,
      "loss": 0.0256,
      "step": 1178
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11003972490068775,
      "grad_norm": 0.1288004070520401,
      "learning_rate": 2e-07,
      "loss": 0.0335,
      "step": 1179
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11013305800068833,
      "grad_norm": 0.1405888944864273,
      "learning_rate": 2e-07,
      "loss": 0.0984,
      "step": 1180
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11022639110068891,
      "grad_norm": 0.11853551864624023,
      "learning_rate": 2e-07,
      "loss": 0.0486,
      "step": 1181
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1103197242006895,
      "grad_norm": 0.12615849077701569,
      "learning_rate": 2e-07,
      "loss": 0.0701,
      "step": 1182
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11041305730069008,
      "grad_norm": 0.12660016119480133,
      "learning_rate": 2e-07,
      "loss": 0.022,
      "step": 1183
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11050639040069067,
      "grad_norm": 0.13127849996089935,
      "learning_rate": 2e-07,
      "loss": 0.0544,
      "step": 1184
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0126953125,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4086.0,
      "completions/mean_length": 608.6767578125,
      "completions/mean_terminated_length": 563.8348388671875,
      "completions/min_length": 95.0,
      "completions/min_terminated_length": 95.0,
      "epoch": 0.11059972350069125,
      "grad_norm": 0.13055351376533508,
      "learning_rate": 2e-07,
      "loss": -0.0166,
      "num_tokens": 818799714.0,
      "reward": 0.5965750813484192,
      "reward_std": 0.18743781745433807,
      "rewards/simpleverify_reward/mean": 0.5965750813484192,
      "rewards/simpleverify_reward/std": 0.49058884382247925,
      "step": 1185
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11069305660069183,
      "grad_norm": 0.12696699798107147,
      "learning_rate": 2e-07,
      "loss": 0.0274,
      "step": 1186
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11078638970069242,
      "grad_norm": 0.1355103999376297,
      "learning_rate": 2e-07,
      "loss": -0.0022,
      "step": 1187
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.110879722800693,
      "grad_norm": 0.12477428466081619,
      "learning_rate": 2e-07,
      "loss": -0.0208,
      "step": 1188
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11097305590069358,
      "grad_norm": 0.12387882173061371,
      "learning_rate": 2e-07,
      "loss": 0.0336,
      "step": 1189
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11106638900069417,
      "grad_norm": 0.12007330358028412,
      "learning_rate": 2e-07,
      "loss": 0.0144,
      "step": 1190
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11115972210069475,
      "grad_norm": 0.12894201278686523,
      "learning_rate": 2e-07,
      "loss": -0.0123,
      "step": 1191
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11125305520069532,
      "grad_norm": 0.11981654167175293,
      "learning_rate": 2e-07,
      "loss": 0.0037,
      "step": 1192
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11134638830069592,
      "grad_norm": 0.1369446963071823,
      "learning_rate": 2e-07,
      "loss": 0.0503,
      "step": 1193
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1114397214006965,
      "grad_norm": 0.12924689054489136,
      "learning_rate": 2e-07,
      "loss": 0.038,
      "step": 1194
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11153305450069709,
      "grad_norm": 0.116915762424469,
      "learning_rate": 2e-07,
      "loss": 0.1079,
      "step": 1195
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11162638760069767,
      "grad_norm": 0.11777377128601074,
      "learning_rate": 2e-07,
      "loss": -0.0137,
      "step": 1196
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11171972070069824,
      "grad_norm": 0.1275106519460678,
      "learning_rate": 2e-07,
      "loss": 0.0412,
      "step": 1197
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11181305380069884,
      "grad_norm": 0.13204318284988403,
      "learning_rate": 2e-07,
      "loss": 0.0141,
      "step": 1198
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11190638690069941,
      "grad_norm": 0.1237158551812172,
      "learning_rate": 2e-07,
      "loss": 0.0397,
      "step": 1199
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11199972000069999,
      "grad_norm": 0.13569581508636475,
      "learning_rate": 2e-07,
      "loss": 0.044,
      "step": 1200
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11209305310070058,
      "grad_norm": 0.11376943439245224,
      "learning_rate": 2e-07,
      "loss": 0.0012,
      "step": 1201
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11218638620070116,
      "grad_norm": 0.12897680699825287,
      "learning_rate": 2e-07,
      "loss": 0.0278,
      "step": 1202
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11227971930070174,
      "grad_norm": 0.12899047136306763,
      "learning_rate": 2e-07,
      "loss": 0.0144,
      "step": 1203
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11237305240070233,
      "grad_norm": 0.14368315041065216,
      "learning_rate": 2e-07,
      "loss": 0.0578,
      "step": 1204
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11246638550070291,
      "grad_norm": 0.12675721943378448,
      "learning_rate": 2e-07,
      "loss": 0.0301,
      "step": 1205
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1125597186007035,
      "grad_norm": 0.14048494398593903,
      "learning_rate": 2e-07,
      "loss": 0.0417,
      "step": 1206
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11265305170070408,
      "grad_norm": 0.1362312138080597,
      "learning_rate": 2e-07,
      "loss": 0.1037,
      "step": 1207
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11274638480070466,
      "grad_norm": 0.12211427837610245,
      "learning_rate": 2e-07,
      "loss": 0.019,
      "step": 1208
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11283971790070525,
      "grad_norm": 0.13619595766067505,
      "learning_rate": 2e-07,
      "loss": -0.0398,
      "step": 1209
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11293305100070583,
      "grad_norm": 0.11813727021217346,
      "learning_rate": 2e-07,
      "loss": 0.0109,
      "step": 1210
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11302638410070641,
      "grad_norm": 0.17518094182014465,
      "learning_rate": 2e-07,
      "loss": 0.0258,
      "step": 1211
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.113119717200707,
      "grad_norm": 0.1293713003396988,
      "learning_rate": 2e-07,
      "loss": 0.0342,
      "step": 1212
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11321305030070758,
      "grad_norm": 0.12582015991210938,
      "learning_rate": 2e-07,
      "loss": 0.0084,
      "step": 1213
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11330638340070817,
      "grad_norm": 0.13867539167404175,
      "learning_rate": 2e-07,
      "loss": 0.0422,
      "step": 1214
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11339971650070875,
      "grad_norm": 0.13374628126621246,
      "learning_rate": 2e-07,
      "loss": 0.0642,
      "step": 1215
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11349304960070933,
      "grad_norm": 0.12595230340957642,
      "learning_rate": 2e-07,
      "loss": 0.0163,
      "step": 1216
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11358638270070992,
      "grad_norm": 0.1340281218290329,
      "learning_rate": 2e-07,
      "loss": -0.005,
      "step": 1217
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1136797158007105,
      "grad_norm": 0.12060722708702087,
      "learning_rate": 2e-07,
      "loss": 0.0036,
      "step": 1218
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11377304890071108,
      "grad_norm": 0.12858417630195618,
      "learning_rate": 2e-07,
      "loss": 0.0018,
      "step": 1219
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11386638200071167,
      "grad_norm": 0.13602882623672485,
      "learning_rate": 2e-07,
      "loss": 0.0213,
      "step": 1220
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11395971510071225,
      "grad_norm": 0.13078676164150238,
      "learning_rate": 2e-07,
      "loss": 0.0145,
      "step": 1221
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11405304820071283,
      "grad_norm": 0.1232333853840828,
      "learning_rate": 2e-07,
      "loss": 0.0592,
      "step": 1222
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11414638130071342,
      "grad_norm": 0.10972514748573303,
      "learning_rate": 2e-07,
      "loss": 0.0369,
      "step": 1223
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.114239714400714,
      "grad_norm": 0.1329655796289444,
      "learning_rate": 2e-07,
      "loss": 0.0343,
      "step": 1224
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11433304750071459,
      "grad_norm": 0.1647261679172516,
      "learning_rate": 2e-07,
      "loss": 0.0282,
      "step": 1225
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11442638060071517,
      "grad_norm": 0.23968738317489624,
      "learning_rate": 2e-07,
      "loss": 0.0271,
      "step": 1226
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11451971370071574,
      "grad_norm": 0.12331417948007584,
      "learning_rate": 2e-07,
      "loss": 0.0584,
      "step": 1227
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11461304680071634,
      "grad_norm": 0.12272704392671585,
      "learning_rate": 2e-07,
      "loss": 0.0427,
      "step": 1228
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11470637990071691,
      "grad_norm": 0.12608398497104645,
      "learning_rate": 2e-07,
      "loss": 0.015,
      "step": 1229
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1147997130007175,
      "grad_norm": 0.12190182507038116,
      "learning_rate": 2e-07,
      "loss": 0.07,
      "step": 1230
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11489304610071809,
      "grad_norm": 0.1147938147187233,
      "learning_rate": 2e-07,
      "loss": 0.0108,
      "step": 1231
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11498637920071866,
      "grad_norm": 0.11709746718406677,
      "learning_rate": 2e-07,
      "loss": -0.0009,
      "step": 1232
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11507971230071924,
      "grad_norm": 0.1322403997182846,
      "learning_rate": 2e-07,
      "loss": 0.0431,
      "step": 1233
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11517304540071983,
      "grad_norm": 0.1292407363653183,
      "learning_rate": 2e-07,
      "loss": 0.0141,
      "step": 1234
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11526637850072041,
      "grad_norm": 0.2323407083749771,
      "learning_rate": 2e-07,
      "loss": 0.0406,
      "step": 1235
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.115359711600721,
      "grad_norm": 0.13938459753990173,
      "learning_rate": 2e-07,
      "loss": -0.0354,
      "step": 1236
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11545304470072158,
      "grad_norm": 0.12293282151222229,
      "learning_rate": 2e-07,
      "loss": 0.0237,
      "step": 1237
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11554637780072216,
      "grad_norm": 0.1341751217842102,
      "learning_rate": 2e-07,
      "loss": 0.0267,
      "step": 1238
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11563971090072275,
      "grad_norm": 0.12490011006593704,
      "learning_rate": 2e-07,
      "loss": 0.045,
      "step": 1239
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11573304400072333,
      "grad_norm": 0.1277049034833908,
      "learning_rate": 2e-07,
      "loss": 0.0421,
      "step": 1240
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11582637710072391,
      "grad_norm": 0.1135387122631073,
      "learning_rate": 2e-07,
      "loss": 0.081,
      "step": 1241
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1159197102007245,
      "grad_norm": 0.11486272513866425,
      "learning_rate": 2e-07,
      "loss": 0.0155,
      "step": 1242
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11601304330072508,
      "grad_norm": 0.13372312486171722,
      "learning_rate": 2e-07,
      "loss": 0.0285,
      "step": 1243
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11610637640072566,
      "grad_norm": 0.1335843801498413,
      "learning_rate": 2e-07,
      "loss": 0.0242,
      "step": 1244
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11619970950072625,
      "grad_norm": 0.1441110223531723,
      "learning_rate": 2e-07,
      "loss": 0.0589,
      "step": 1245
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11629304260072683,
      "grad_norm": 0.13124550879001617,
      "learning_rate": 2e-07,
      "loss": 0.0586,
      "step": 1246
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11638637570072742,
      "grad_norm": 0.12812970578670502,
      "learning_rate": 2e-07,
      "loss": 0.0505,
      "step": 1247
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.116479708800728,
      "grad_norm": 0.14105205237865448,
      "learning_rate": 2e-07,
      "loss": 0.0411,
      "step": 1248
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.011858258928571397,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4096.0,
      "completions/mean_length": 608.3168334960938,
      "completions/mean_terminated_length": 566.462646484375,
      "completions/min_length": 92.0,
      "completions/min_terminated_length": 92.0,
      "epoch": 0.11657304190072858,
      "grad_norm": 0.12530314922332764,
      "learning_rate": 2e-07,
      "loss": 0.0192,
      "num_tokens": 859315523.0,
      "reward": 0.5994175672531128,
      "reward_std": 0.1901005059480667,
      "rewards/simpleverify_reward/mean": 0.5994175672531128,
      "rewards/simpleverify_reward/std": 0.4900206923484802,
      "step": 1249
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11666637500072917,
      "grad_norm": 0.13039752840995789,
      "learning_rate": 2e-07,
      "loss": 0.001,
      "step": 1250
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11675970810072975,
      "grad_norm": 0.14744889736175537,
      "learning_rate": 2e-07,
      "loss": 0.0467,
      "step": 1251
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11685304120073033,
      "grad_norm": 0.14962077140808105,
      "learning_rate": 2e-07,
      "loss": 0.0156,
      "step": 1252
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11694637430073092,
      "grad_norm": 0.1135014146566391,
      "learning_rate": 2e-07,
      "loss": 0.0235,
      "step": 1253
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1170397074007315,
      "grad_norm": 0.12785571813583374,
      "learning_rate": 2e-07,
      "loss": -0.0171,
      "step": 1254
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11713304050073207,
      "grad_norm": 0.1328587681055069,
      "learning_rate": 2e-07,
      "loss": 0.0297,
      "step": 1255
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11722637360073267,
      "grad_norm": 0.12974928319454193,
      "learning_rate": 2e-07,
      "loss": 0.0488,
      "step": 1256
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11731970670073325,
      "grad_norm": 0.13236884772777557,
      "learning_rate": 2e-07,
      "loss": -0.0041,
      "step": 1257
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11741303980073384,
      "grad_norm": 0.12275372445583344,
      "learning_rate": 2e-07,
      "loss": 0.0284,
      "step": 1258
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11750637290073442,
      "grad_norm": 0.12199831753969193,
      "learning_rate": 2e-07,
      "loss": 0.0517,
      "step": 1259
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.117599706000735,
      "grad_norm": 0.10939118266105652,
      "learning_rate": 2e-07,
      "loss": 0.0514,
      "step": 1260
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11769303910073559,
      "grad_norm": 0.1254461705684662,
      "learning_rate": 2e-07,
      "loss": 0.0361,
      "step": 1261
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11778637220073616,
      "grad_norm": 0.14978745579719543,
      "learning_rate": 2e-07,
      "loss": 0.0263,
      "step": 1262
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11787970530073674,
      "grad_norm": 0.13184259831905365,
      "learning_rate": 2e-07,
      "loss": -0.0347,
      "step": 1263
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11797303840073733,
      "grad_norm": 0.1274329572916031,
      "learning_rate": 2e-07,
      "loss": 0.0283,
      "step": 1264
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11806637150073791,
      "grad_norm": 0.12235521525144577,
      "learning_rate": 2e-07,
      "loss": 0.0188,
      "step": 1265
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1181597046007385,
      "grad_norm": 0.12771612405776978,
      "learning_rate": 2e-07,
      "loss": 0.0374,
      "step": 1266
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11825303770073908,
      "grad_norm": 0.14621379971504211,
      "learning_rate": 2e-07,
      "loss": 0.0711,
      "step": 1267
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11834637080073966,
      "grad_norm": 0.12454205751419067,
      "learning_rate": 2e-07,
      "loss": 0.0449,
      "step": 1268
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11843970390074025,
      "grad_norm": 0.12909890711307526,
      "learning_rate": 2e-07,
      "loss": 0.0406,
      "step": 1269
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11853303700074083,
      "grad_norm": 0.1296859085559845,
      "learning_rate": 2e-07,
      "loss": 0.0226,
      "step": 1270
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11862637010074141,
      "grad_norm": 0.12741021811962128,
      "learning_rate": 2e-07,
      "loss": 0.0523,
      "step": 1271
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.118719703200742,
      "grad_norm": 0.1277891993522644,
      "learning_rate": 2e-07,
      "loss": -0.017,
      "step": 1272
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11881303630074258,
      "grad_norm": 0.13933908939361572,
      "learning_rate": 2e-07,
      "loss": 0.0436,
      "step": 1273
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11890636940074316,
      "grad_norm": 0.12748397886753082,
      "learning_rate": 2e-07,
      "loss": 0.0197,
      "step": 1274
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11899970250074375,
      "grad_norm": 0.1312733292579651,
      "learning_rate": 2e-07,
      "loss": 0.1026,
      "step": 1275
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11909303560074433,
      "grad_norm": 0.19052410125732422,
      "learning_rate": 2e-07,
      "loss": 0.0519,
      "step": 1276
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11918636870074492,
      "grad_norm": 0.1322513371706009,
      "learning_rate": 2e-07,
      "loss": -0.0025,
      "step": 1277
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1192797018007455,
      "grad_norm": 0.12174084037542343,
      "learning_rate": 2e-07,
      "loss": 0.0332,
      "step": 1278
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11937303490074608,
      "grad_norm": 0.12063100934028625,
      "learning_rate": 2e-07,
      "loss": 0.0359,
      "step": 1279
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11946636800074667,
      "grad_norm": 0.1366986632347107,
      "learning_rate": 2e-07,
      "loss": 0.0171,
      "step": 1280
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11955970110074725,
      "grad_norm": 0.13286036252975464,
      "learning_rate": 2e-07,
      "loss": 0.0453,
      "step": 1281
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11965303420074783,
      "grad_norm": 0.12996463477611542,
      "learning_rate": 2e-07,
      "loss": 0.0296,
      "step": 1282
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11974636730074842,
      "grad_norm": 0.13416169583797455,
      "learning_rate": 2e-07,
      "loss": 0.0216,
      "step": 1283
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.119839700400749,
      "grad_norm": 0.12199593335390091,
      "learning_rate": 2e-07,
      "loss": 0.0118,
      "step": 1284
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.11993303350074958,
      "grad_norm": 0.12001723796129227,
      "learning_rate": 2e-07,
      "loss": 0.0646,
      "step": 1285
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12002636660075017,
      "grad_norm": 0.12678152322769165,
      "learning_rate": 2e-07,
      "loss": -0.0011,
      "step": 1286
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12011969970075075,
      "grad_norm": 0.12558075785636902,
      "learning_rate": 2e-07,
      "loss": 0.0131,
      "step": 1287
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12021303280075134,
      "grad_norm": 0.13155266642570496,
      "learning_rate": 2e-07,
      "loss": 0.0163,
      "step": 1288
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12030636590075192,
      "grad_norm": 0.13256101310253143,
      "learning_rate": 2e-07,
      "loss": 0.0163,
      "step": 1289
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1203996990007525,
      "grad_norm": 0.15142793953418732,
      "learning_rate": 2e-07,
      "loss": 0.0032,
      "step": 1290
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12049303210075309,
      "grad_norm": 0.13562975823879242,
      "learning_rate": 2e-07,
      "loss": 0.0168,
      "step": 1291
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12058636520075366,
      "grad_norm": 0.1157451644539833,
      "learning_rate": 2e-07,
      "loss": 0.0323,
      "step": 1292
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12067969830075424,
      "grad_norm": 0.12735344469547272,
      "learning_rate": 2e-07,
      "loss": 0.0513,
      "step": 1293
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12077303140075484,
      "grad_norm": 0.1280812919139862,
      "learning_rate": 2e-07,
      "loss": 0.0328,
      "step": 1294
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12086636450075541,
      "grad_norm": 0.142857164144516,
      "learning_rate": 2e-07,
      "loss": 0.0534,
      "step": 1295
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12095969760075599,
      "grad_norm": 0.13524094223976135,
      "learning_rate": 2e-07,
      "loss": 0.0192,
      "step": 1296
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12105303070075658,
      "grad_norm": 0.13018885254859924,
      "learning_rate": 2e-07,
      "loss": 0.052,
      "step": 1297
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12114636380075716,
      "grad_norm": 0.12639300525188446,
      "learning_rate": 2e-07,
      "loss": -0.0129,
      "step": 1298
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12123969690075775,
      "grad_norm": 0.13235361874103546,
      "learning_rate": 2e-07,
      "loss": -0.0366,
      "step": 1299
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12133303000075833,
      "grad_norm": 0.14917443692684174,
      "learning_rate": 2e-07,
      "loss": 0.0161,
      "step": 1300
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12142636310075891,
      "grad_norm": 0.14884494245052338,
      "learning_rate": 2e-07,
      "loss": 0.0409,
      "step": 1301
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1215196962007595,
      "grad_norm": 0.15109430253505707,
      "learning_rate": 2e-07,
      "loss": 0.0272,
      "step": 1302
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12161302930076008,
      "grad_norm": 0.13050200045108795,
      "learning_rate": 2e-07,
      "loss": 0.0099,
      "step": 1303
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12170636240076066,
      "grad_norm": 0.11470898985862732,
      "learning_rate": 2e-07,
      "loss": 0.0107,
      "step": 1304
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12179969550076125,
      "grad_norm": 0.14495515823364258,
      "learning_rate": 2e-07,
      "loss": 0.0504,
      "step": 1305
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12189302860076183,
      "grad_norm": 0.1229834333062172,
      "learning_rate": 2e-07,
      "loss": -0.014,
      "step": 1306
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12198636170076241,
      "grad_norm": 0.1366640031337738,
      "learning_rate": 2e-07,
      "loss": 0.0114,
      "step": 1307
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.122079694800763,
      "grad_norm": 0.13780227303504944,
      "learning_rate": 2e-07,
      "loss": 0.0231,
      "step": 1308
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12217302790076358,
      "grad_norm": 0.1299963891506195,
      "learning_rate": 2e-07,
      "loss": -0.0283,
      "step": 1309
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12226636100076417,
      "grad_norm": 0.12133779376745224,
      "learning_rate": 2e-07,
      "loss": -0.0305,
      "step": 1310
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12235969410076475,
      "grad_norm": 0.12540125846862793,
      "learning_rate": 2e-07,
      "loss": 0.0219,
      "step": 1311
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12245302720076533,
      "grad_norm": 0.13978445529937744,
      "learning_rate": 2e-07,
      "loss": 0.0207,
      "step": 1312
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.012555803571428603,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4096.0,
      "completions/mean_length": 607.9358520507812,
      "completions/mean_terminated_length": 563.5834350585938,
      "completions/min_length": 78.0,
      "completions/min_terminated_length": 78.0,
      "epoch": 0.12254636030076592,
      "grad_norm": 0.11809849739074707,
      "learning_rate": 2e-07,
      "loss": 0.0165,
      "num_tokens": 899854737.0,
      "reward": 0.5961216688156128,
      "reward_std": 0.19023947417736053,
      "rewards/simpleverify_reward/mean": 0.5961216688156128,
      "rewards/simpleverify_reward/std": 0.49067792296409607,
      "step": 1313
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1226396934007665,
      "grad_norm": 0.12058080732822418,
      "learning_rate": 2e-07,
      "loss": 0.0601,
      "step": 1314
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12273302650076708,
      "grad_norm": 0.125620037317276,
      "learning_rate": 2e-07,
      "loss": 0.016,
      "step": 1315
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12282635960076767,
      "grad_norm": 0.12974177300930023,
      "learning_rate": 2e-07,
      "loss": 0.0512,
      "step": 1316
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12291969270076825,
      "grad_norm": 0.1465127170085907,
      "learning_rate": 2e-07,
      "loss": 0.0096,
      "step": 1317
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12301302580076884,
      "grad_norm": 0.14485350251197815,
      "learning_rate": 2e-07,
      "loss": -0.0101,
      "step": 1318
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12310635890076942,
      "grad_norm": 0.1286245733499527,
      "learning_rate": 2e-07,
      "loss": 0.0203,
      "step": 1319
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12319969200077,
      "grad_norm": 0.12529346346855164,
      "learning_rate": 2e-07,
      "loss": 0.0662,
      "step": 1320
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12329302510077059,
      "grad_norm": 0.1323009878396988,
      "learning_rate": 2e-07,
      "loss": -0.0042,
      "step": 1321
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12338635820077117,
      "grad_norm": 0.11834096163511276,
      "learning_rate": 2e-07,
      "loss": 0.0508,
      "step": 1322
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12347969130077174,
      "grad_norm": 0.12331033498048782,
      "learning_rate": 2e-07,
      "loss": 0.0369,
      "step": 1323
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12357302440077234,
      "grad_norm": 0.12092196941375732,
      "learning_rate": 2e-07,
      "loss": 0.0079,
      "step": 1324
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12366635750077291,
      "grad_norm": 0.13256673514842987,
      "learning_rate": 2e-07,
      "loss": 0.0218,
      "step": 1325
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12375969060077349,
      "grad_norm": 0.13517709076404572,
      "learning_rate": 2e-07,
      "loss": 0.1028,
      "step": 1326
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12385302370077408,
      "grad_norm": 0.12170856446027756,
      "learning_rate": 2e-07,
      "loss": 0.0972,
      "step": 1327
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12394635680077466,
      "grad_norm": 0.12465059012174606,
      "learning_rate": 2e-07,
      "loss": 0.0062,
      "step": 1328
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12403968990077525,
      "grad_norm": 0.13149112462997437,
      "learning_rate": 2e-07,
      "loss": 0.0087,
      "step": 1329
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12413302300077583,
      "grad_norm": 0.11944835633039474,
      "learning_rate": 2e-07,
      "loss": 0.015,
      "step": 1330
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12422635610077641,
      "grad_norm": 0.12846983969211578,
      "learning_rate": 2e-07,
      "loss": -0.0237,
      "step": 1331
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.124319689200777,
      "grad_norm": 0.12928855419158936,
      "learning_rate": 2e-07,
      "loss": -0.0063,
      "step": 1332
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12441302230077758,
      "grad_norm": 0.12158067524433136,
      "learning_rate": 2e-07,
      "loss": 0.0373,
      "step": 1333
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12450635540077816,
      "grad_norm": 0.1204054057598114,
      "learning_rate": 2e-07,
      "loss": -0.0123,
      "step": 1334
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12459968850077875,
      "grad_norm": 0.12858541309833527,
      "learning_rate": 2e-07,
      "loss": 0.047,
      "step": 1335
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12469302160077933,
      "grad_norm": 0.12837396562099457,
      "learning_rate": 2e-07,
      "loss": 0.0171,
      "step": 1336
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12478635470077991,
      "grad_norm": 0.1236601248383522,
      "learning_rate": 2e-07,
      "loss": 0.0411,
      "step": 1337
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1248796878007805,
      "grad_norm": 0.133812814950943,
      "learning_rate": 2e-07,
      "loss": 0.0218,
      "step": 1338
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12497302090078108,
      "grad_norm": 0.12431017309427261,
      "learning_rate": 2e-07,
      "loss": 0.0609,
      "step": 1339
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12506635400078167,
      "grad_norm": 0.1297278255224228,
      "learning_rate": 2e-07,
      "loss": 0.0609,
      "step": 1340
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12515968710078224,
      "grad_norm": 0.13260437548160553,
      "learning_rate": 2e-07,
      "loss": 0.023,
      "step": 1341
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12525302020078283,
      "grad_norm": 0.13998012244701385,
      "learning_rate": 2e-07,
      "loss": 0.0367,
      "step": 1342
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12534635330078342,
      "grad_norm": 0.13109180331230164,
      "learning_rate": 2e-07,
      "loss": 0.0317,
      "step": 1343
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.125439686400784,
      "grad_norm": 0.1368030607700348,
      "learning_rate": 2e-07,
      "loss": 0.0573,
      "step": 1344
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12553301950078458,
      "grad_norm": 0.13046666979789734,
      "learning_rate": 2e-07,
      "loss": 0.0615,
      "step": 1345
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12562635260078517,
      "grad_norm": 0.12309705466032028,
      "learning_rate": 2e-07,
      "loss": 0.013,
      "step": 1346
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12571968570078576,
      "grad_norm": 0.14114674925804138,
      "learning_rate": 2e-07,
      "loss": 0.0253,
      "step": 1347
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12581301880078632,
      "grad_norm": 0.1223163977265358,
      "learning_rate": 2e-07,
      "loss": 0.0312,
      "step": 1348
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12590635190078692,
      "grad_norm": 0.1385715752840042,
      "learning_rate": 2e-07,
      "loss": -0.0011,
      "step": 1349
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1259996850007875,
      "grad_norm": 0.12719270586967468,
      "learning_rate": 2e-07,
      "loss": -0.0216,
      "step": 1350
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12609301810078807,
      "grad_norm": 0.1360722929239273,
      "learning_rate": 2e-07,
      "loss": 0.0301,
      "step": 1351
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12618635120078867,
      "grad_norm": 0.1249774917960167,
      "learning_rate": 2e-07,
      "loss": 0.0363,
      "step": 1352
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12627968430078926,
      "grad_norm": 0.12800733745098114,
      "learning_rate": 2e-07,
      "loss": 0.0119,
      "step": 1353
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12637301740078982,
      "grad_norm": 0.17600104212760925,
      "learning_rate": 2e-07,
      "loss": 0.0151,
      "step": 1354
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12646635050079041,
      "grad_norm": 0.15301010012626648,
      "learning_rate": 2e-07,
      "loss": 0.0049,
      "step": 1355
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.126559683600791,
      "grad_norm": 0.1212579682469368,
      "learning_rate": 2e-07,
      "loss": 0.0208,
      "step": 1356
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12665301670079157,
      "grad_norm": 0.13823603093624115,
      "learning_rate": 2e-07,
      "loss": -0.0217,
      "step": 1357
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12674634980079216,
      "grad_norm": 0.12493650615215302,
      "learning_rate": 2e-07,
      "loss": 0.005,
      "step": 1358
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12683968290079276,
      "grad_norm": 0.13408571481704712,
      "learning_rate": 2e-07,
      "loss": 0.068,
      "step": 1359
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12693301600079332,
      "grad_norm": 0.14997972548007965,
      "learning_rate": 2e-07,
      "loss": 0.0373,
      "step": 1360
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1270263491007939,
      "grad_norm": 0.13704782724380493,
      "learning_rate": 2e-07,
      "loss": 0.0693,
      "step": 1361
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1271196822007945,
      "grad_norm": 0.12894350290298462,
      "learning_rate": 2e-07,
      "loss": 0.0351,
      "step": 1362
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12721301530079507,
      "grad_norm": 0.14157943427562714,
      "learning_rate": 2e-07,
      "loss": 0.022,
      "step": 1363
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12730634840079566,
      "grad_norm": 0.13020886480808258,
      "learning_rate": 2e-07,
      "loss": 0.0376,
      "step": 1364
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12739968150079625,
      "grad_norm": 0.11284741759300232,
      "learning_rate": 2e-07,
      "loss": 0.049,
      "step": 1365
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12749301460079684,
      "grad_norm": 0.12828324735164642,
      "learning_rate": 2e-07,
      "loss": 0.0226,
      "step": 1366
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1275863477007974,
      "grad_norm": 0.14629149436950684,
      "learning_rate": 2e-07,
      "loss": 0.0188,
      "step": 1367
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.127679680800798,
      "grad_norm": 0.12480209767818451,
      "learning_rate": 2e-07,
      "loss": -0.0348,
      "step": 1368
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1277730139007986,
      "grad_norm": 0.14701150357723236,
      "learning_rate": 2e-07,
      "loss": 0.0068,
      "step": 1369
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12786634700079916,
      "grad_norm": 0.12233898043632507,
      "learning_rate": 2e-07,
      "loss": -0.0118,
      "step": 1370
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12795968010079975,
      "grad_norm": 0.13516168296337128,
      "learning_rate": 2e-07,
      "loss": 0.0266,
      "step": 1371
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12805301320080034,
      "grad_norm": 0.14430968463420868,
      "learning_rate": 2e-07,
      "loss": 0.0487,
      "step": 1372
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1281463463008009,
      "grad_norm": 0.12676605582237244,
      "learning_rate": 2e-07,
      "loss": 0.0945,
      "step": 1373
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1282396794008015,
      "grad_norm": 0.1392187625169754,
      "learning_rate": 2e-07,
      "loss": 0.0426,
      "step": 1374
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1283330125008021,
      "grad_norm": 0.12715740501880646,
      "learning_rate": 2e-07,
      "loss": 0.0379,
      "step": 1375
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12842634560080265,
      "grad_norm": 0.13816578686237335,
      "learning_rate": 2e-07,
      "loss": 0.0257,
      "step": 1376
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.013044084821428603,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4091.0,
      "completions/mean_length": 614.19140625,
      "completions/mean_terminated_length": 568.1741333007812,
      "completions/min_length": 92.0,
      "completions/min_terminated_length": 92.0,
      "epoch": 0.12851967870080325,
      "grad_norm": 0.12122588604688644,
      "learning_rate": 2e-07,
      "loss": 0.0611,
      "num_tokens": 940763185.0,
      "reward": 0.5921630859375,
      "reward_std": 0.18556271493434906,
      "rewards/simpleverify_reward/mean": 0.5921630859375,
      "rewards/simpleverify_reward/std": 0.49143680930137634,
      "step": 1377
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12861301180080384,
      "grad_norm": 0.1205112412571907,
      "learning_rate": 2e-07,
      "loss": 0.0333,
      "step": 1378
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1287063449008044,
      "grad_norm": 0.12911224365234375,
      "learning_rate": 2e-07,
      "loss": 0.073,
      "step": 1379
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.128799678000805,
      "grad_norm": 0.12581101059913635,
      "learning_rate": 2e-07,
      "loss": 0.0222,
      "step": 1380
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1288930111008056,
      "grad_norm": 0.11776664108037949,
      "learning_rate": 2e-07,
      "loss": 0.0123,
      "step": 1381
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12898634420080615,
      "grad_norm": 0.11414356529712677,
      "learning_rate": 2e-07,
      "loss": 0.0512,
      "step": 1382
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12907967730080674,
      "grad_norm": 0.1338527351617813,
      "learning_rate": 2e-07,
      "loss": 0.0084,
      "step": 1383
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12917301040080734,
      "grad_norm": 0.1056576818227768,
      "learning_rate": 2e-07,
      "loss": 0.0722,
      "step": 1384
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1292663435008079,
      "grad_norm": 0.12868031859397888,
      "learning_rate": 2e-07,
      "loss": 0.0189,
      "step": 1385
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1293596766008085,
      "grad_norm": 0.11210151761770248,
      "learning_rate": 2e-07,
      "loss": 0.0338,
      "step": 1386
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12945300970080909,
      "grad_norm": 0.1337253302335739,
      "learning_rate": 2e-07,
      "loss": -0.0242,
      "step": 1387
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12954634280080968,
      "grad_norm": 0.13245192170143127,
      "learning_rate": 2e-07,
      "loss": -0.009,
      "step": 1388
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12963967590081024,
      "grad_norm": 0.1296364814043045,
      "learning_rate": 2e-07,
      "loss": 0.041,
      "step": 1389
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12973300900081083,
      "grad_norm": 0.11979067325592041,
      "learning_rate": 2e-07,
      "loss": 0.0326,
      "step": 1390
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.12982634210081143,
      "grad_norm": 0.11506450176239014,
      "learning_rate": 2e-07,
      "loss": 0.0056,
      "step": 1391
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.129919675200812,
      "grad_norm": 0.12648504972457886,
      "learning_rate": 2e-07,
      "loss": 0.0118,
      "step": 1392
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13001300830081258,
      "grad_norm": 0.1077185645699501,
      "learning_rate": 2e-07,
      "loss": 0.0581,
      "step": 1393
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13010634140081317,
      "grad_norm": 0.14138028025627136,
      "learning_rate": 2e-07,
      "loss": 0.0433,
      "step": 1394
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13019967450081374,
      "grad_norm": 0.1270858645439148,
      "learning_rate": 2e-07,
      "loss": 0.017,
      "step": 1395
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13029300760081433,
      "grad_norm": 0.12222479283809662,
      "learning_rate": 2e-07,
      "loss": -0.0203,
      "step": 1396
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13038634070081492,
      "grad_norm": 0.1286737322807312,
      "learning_rate": 2e-07,
      "loss": 0.0113,
      "step": 1397
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1304796738008155,
      "grad_norm": 0.12092230468988419,
      "learning_rate": 2e-07,
      "loss": 0.0144,
      "step": 1398
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13057300690081608,
      "grad_norm": 0.1365211457014084,
      "learning_rate": 2e-07,
      "loss": 0.0064,
      "step": 1399
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13066634000081667,
      "grad_norm": 0.12843672931194305,
      "learning_rate": 2e-07,
      "loss": 0.0236,
      "step": 1400
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13075967310081724,
      "grad_norm": 0.11588891595602036,
      "learning_rate": 2e-07,
      "loss": 0.0152,
      "step": 1401
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13085300620081783,
      "grad_norm": 0.1260339617729187,
      "learning_rate": 2e-07,
      "loss": 0.0571,
      "step": 1402
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13094633930081842,
      "grad_norm": 0.1260555386543274,
      "learning_rate": 2e-07,
      "loss": 0.0285,
      "step": 1403
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13103967240081899,
      "grad_norm": 0.13749821484088898,
      "learning_rate": 2e-07,
      "loss": -0.0126,
      "step": 1404
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13113300550081958,
      "grad_norm": 0.12537643313407898,
      "learning_rate": 2e-07,
      "loss": 0.0342,
      "step": 1405
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13122633860082017,
      "grad_norm": 0.13549663126468658,
      "learning_rate": 2e-07,
      "loss": -0.0033,
      "step": 1406
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13131967170082076,
      "grad_norm": 0.1296100616455078,
      "learning_rate": 2e-07,
      "loss": 0.0417,
      "step": 1407
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13141300480082133,
      "grad_norm": 0.11951570212841034,
      "learning_rate": 2e-07,
      "loss": 0.0588,
      "step": 1408
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13150633790082192,
      "grad_norm": 0.12128618359565735,
      "learning_rate": 2e-07,
      "loss": 0.0182,
      "step": 1409
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1315996710008225,
      "grad_norm": 0.1259547919034958,
      "learning_rate": 2e-07,
      "loss": -0.0133,
      "step": 1410
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13169300410082307,
      "grad_norm": 0.12305407971143723,
      "learning_rate": 2e-07,
      "loss": 0.0231,
      "step": 1411
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13178633720082367,
      "grad_norm": 0.12852798402309418,
      "learning_rate": 2e-07,
      "loss": -0.0356,
      "step": 1412
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13187967030082426,
      "grad_norm": 0.12451092898845673,
      "learning_rate": 2e-07,
      "loss": 0.0249,
      "step": 1413
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13197300340082482,
      "grad_norm": 0.12386109679937363,
      "learning_rate": 2e-07,
      "loss": 0.0406,
      "step": 1414
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13206633650082542,
      "grad_norm": 0.12550488114356995,
      "learning_rate": 2e-07,
      "loss": 0.03,
      "step": 1415
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.132159669600826,
      "grad_norm": 0.13374298810958862,
      "learning_rate": 2e-07,
      "loss": 0.0468,
      "step": 1416
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13225300270082657,
      "grad_norm": 0.13182391226291656,
      "learning_rate": 2e-07,
      "loss": -0.0005,
      "step": 1417
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13234633580082716,
      "grad_norm": 0.12399030476808548,
      "learning_rate": 2e-07,
      "loss": 0.0296,
      "step": 1418
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13243966890082776,
      "grad_norm": 0.13912323117256165,
      "learning_rate": 2e-07,
      "loss": 0.0109,
      "step": 1419
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13253300200082832,
      "grad_norm": 0.1259942352771759,
      "learning_rate": 2e-07,
      "loss": 0.0146,
      "step": 1420
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1326263351008289,
      "grad_norm": 0.12687213718891144,
      "learning_rate": 2e-07,
      "loss": -0.0082,
      "step": 1421
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1327196682008295,
      "grad_norm": 0.13498739898204803,
      "learning_rate": 2e-07,
      "loss": 0.0356,
      "step": 1422
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13281300130083007,
      "grad_norm": 0.14409327507019043,
      "learning_rate": 2e-07,
      "loss": 0.0678,
      "step": 1423
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13290633440083066,
      "grad_norm": 0.1604636013507843,
      "learning_rate": 2e-07,
      "loss": 0.0247,
      "step": 1424
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13299966750083125,
      "grad_norm": 0.13310372829437256,
      "learning_rate": 2e-07,
      "loss": -0.0224,
      "step": 1425
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13309300060083182,
      "grad_norm": 0.12996575236320496,
      "learning_rate": 2e-07,
      "loss": -0.0049,
      "step": 1426
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1331863337008324,
      "grad_norm": 0.1303420513868332,
      "learning_rate": 2e-07,
      "loss": 0.0134,
      "step": 1427
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.133279666800833,
      "grad_norm": 0.14204761385917664,
      "learning_rate": 2e-07,
      "loss": -0.0087,
      "step": 1428
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1333729999008336,
      "grad_norm": 0.13351668417453766,
      "learning_rate": 2e-07,
      "loss": 0.0417,
      "step": 1429
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13346633300083416,
      "grad_norm": 0.13873291015625,
      "learning_rate": 2e-07,
      "loss": 0.0429,
      "step": 1430
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13355966610083475,
      "grad_norm": 0.13662871718406677,
      "learning_rate": 2e-07,
      "loss": 0.0197,
      "step": 1431
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13365299920083534,
      "grad_norm": 0.1312268078327179,
      "learning_rate": 2e-07,
      "loss": 0.0233,
      "step": 1432
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1337463323008359,
      "grad_norm": 0.14710059762001038,
      "learning_rate": 2e-07,
      "loss": 0.0508,
      "step": 1433
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1338396654008365,
      "grad_norm": 0.21976804733276367,
      "learning_rate": 2e-07,
      "loss": 0.0348,
      "step": 1434
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1339329985008371,
      "grad_norm": 0.11967466026544571,
      "learning_rate": 2e-07,
      "loss": 0.0112,
      "step": 1435
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13402633160083766,
      "grad_norm": 0.13249734044075012,
      "learning_rate": 2e-07,
      "loss": 0.0539,
      "step": 1436
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13411966470083825,
      "grad_norm": 0.1204148679971695,
      "learning_rate": 2e-07,
      "loss": 0.0565,
      "step": 1437
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13421299780083884,
      "grad_norm": 0.20165148377418518,
      "learning_rate": 2e-07,
      "loss": 0.0353,
      "step": 1438
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1343063309008394,
      "grad_norm": 0.13655179738998413,
      "learning_rate": 2e-07,
      "loss": 0.0441,
      "step": 1439
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13439966400084,
      "grad_norm": 0.1327507644891739,
      "learning_rate": 2e-07,
      "loss": 0.0275,
      "step": 1440
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01318359375,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4082.0,
      "completions/mean_length": 620.673095703125,
      "completions/mean_terminated_length": 574.24365234375,
      "completions/min_length": 78.0,
      "completions/min_terminated_length": 78.0,
      "epoch": 0.1344929971008406,
      "grad_norm": 0.14014360308647156,
      "learning_rate": 2e-07,
      "loss": 0.0198,
      "num_tokens": 982072539.0,
      "reward": 0.5868443250656128,
      "reward_std": 0.18858738243579865,
      "rewards/simpleverify_reward/mean": 0.5868443250656128,
      "rewards/simpleverify_reward/std": 0.49240460991859436,
      "step": 1441
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13458633020084115,
      "grad_norm": 0.1204172894358635,
      "learning_rate": 2e-07,
      "loss": -0.0031,
      "step": 1442
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13467966330084175,
      "grad_norm": 0.11600620299577713,
      "learning_rate": 2e-07,
      "loss": 0.0314,
      "step": 1443
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13477299640084234,
      "grad_norm": 0.12386095523834229,
      "learning_rate": 2e-07,
      "loss": 0.0401,
      "step": 1444
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1348663295008429,
      "grad_norm": 0.12171697616577148,
      "learning_rate": 2e-07,
      "loss": -0.0037,
      "step": 1445
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1349596626008435,
      "grad_norm": 0.12403026968240738,
      "learning_rate": 2e-07,
      "loss": 0.0421,
      "step": 1446
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1350529957008441,
      "grad_norm": 0.13129982352256775,
      "learning_rate": 2e-07,
      "loss": 0.0079,
      "step": 1447
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13514632880084468,
      "grad_norm": 0.11801241338253021,
      "learning_rate": 2e-07,
      "loss": -0.0167,
      "step": 1448
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13523966190084524,
      "grad_norm": 0.1307920664548874,
      "learning_rate": 2e-07,
      "loss": 0.0087,
      "step": 1449
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13533299500084583,
      "grad_norm": 0.13329008221626282,
      "learning_rate": 2e-07,
      "loss": 0.0434,
      "step": 1450
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13542632810084643,
      "grad_norm": 0.12755557894706726,
      "learning_rate": 2e-07,
      "loss": -0.0231,
      "step": 1451
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.135519661200847,
      "grad_norm": 0.12927190959453583,
      "learning_rate": 2e-07,
      "loss": -0.0311,
      "step": 1452
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13561299430084758,
      "grad_norm": 0.12124588340520859,
      "learning_rate": 2e-07,
      "loss": 0.0296,
      "step": 1453
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13570632740084818,
      "grad_norm": 0.12378164380788803,
      "learning_rate": 2e-07,
      "loss": 0.0516,
      "step": 1454
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13579966050084874,
      "grad_norm": 0.1416238248348236,
      "learning_rate": 2e-07,
      "loss": 0.0491,
      "step": 1455
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13589299360084933,
      "grad_norm": 0.12794120609760284,
      "learning_rate": 2e-07,
      "loss": 0.0571,
      "step": 1456
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13598632670084992,
      "grad_norm": 0.12237957864999771,
      "learning_rate": 2e-07,
      "loss": 0.015,
      "step": 1457
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1360796598008505,
      "grad_norm": 0.13529318571090698,
      "learning_rate": 2e-07,
      "loss": 0.0529,
      "step": 1458
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13617299290085108,
      "grad_norm": 0.1210915669798851,
      "learning_rate": 2e-07,
      "loss": 0.0111,
      "step": 1459
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13626632600085167,
      "grad_norm": 0.14912164211273193,
      "learning_rate": 2e-07,
      "loss": 0.0318,
      "step": 1460
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13635965910085224,
      "grad_norm": 0.13777463138103485,
      "learning_rate": 2e-07,
      "loss": 0.012,
      "step": 1461
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13645299220085283,
      "grad_norm": 0.1252203732728958,
      "learning_rate": 2e-07,
      "loss": 0.0339,
      "step": 1462
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13654632530085342,
      "grad_norm": 0.13650867342948914,
      "learning_rate": 2e-07,
      "loss": -0.0138,
      "step": 1463
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13663965840085399,
      "grad_norm": 0.12399674206972122,
      "learning_rate": 2e-07,
      "loss": 0.0206,
      "step": 1464
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13673299150085458,
      "grad_norm": 0.13404396176338196,
      "learning_rate": 2e-07,
      "loss": 0.0047,
      "step": 1465
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13682632460085517,
      "grad_norm": 0.12794950604438782,
      "learning_rate": 2e-07,
      "loss": 0.0254,
      "step": 1466
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13691965770085573,
      "grad_norm": 0.12589828670024872,
      "learning_rate": 2e-07,
      "loss": 0.0579,
      "step": 1467
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13701299080085633,
      "grad_norm": 0.11313801258802414,
      "learning_rate": 2e-07,
      "loss": 0.0313,
      "step": 1468
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13710632390085692,
      "grad_norm": 0.12220331281423569,
      "learning_rate": 2e-07,
      "loss": 0.0137,
      "step": 1469
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1371996570008575,
      "grad_norm": 0.13046205043792725,
      "learning_rate": 2e-07,
      "loss": 0.0429,
      "step": 1470
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13729299010085808,
      "grad_norm": 0.13688932359218597,
      "learning_rate": 2e-07,
      "loss": -0.0201,
      "step": 1471
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13738632320085867,
      "grad_norm": 0.1379668116569519,
      "learning_rate": 2e-07,
      "loss": -0.0078,
      "step": 1472
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13747965630085926,
      "grad_norm": 0.12323816120624542,
      "learning_rate": 2e-07,
      "loss": 0.0377,
      "step": 1473
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13757298940085982,
      "grad_norm": 0.1078827902674675,
      "learning_rate": 2e-07,
      "loss": 0.0429,
      "step": 1474
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13766632250086042,
      "grad_norm": 0.13498204946517944,
      "learning_rate": 2e-07,
      "loss": 0.0071,
      "step": 1475
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.137759655600861,
      "grad_norm": 0.14558745920658112,
      "learning_rate": 2e-07,
      "loss": 0.0167,
      "step": 1476
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13785298870086157,
      "grad_norm": 0.13273562490940094,
      "learning_rate": 2e-07,
      "loss": 0.025,
      "step": 1477
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13794632180086216,
      "grad_norm": 0.13516178727149963,
      "learning_rate": 2e-07,
      "loss": 0.052,
      "step": 1478
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13803965490086276,
      "grad_norm": 0.16784650087356567,
      "learning_rate": 2e-07,
      "loss": 0.0406,
      "step": 1479
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13813298800086332,
      "grad_norm": 0.1249307319521904,
      "learning_rate": 2e-07,
      "loss": 0.0471,
      "step": 1480
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1382263211008639,
      "grad_norm": 0.14602787792682648,
      "learning_rate": 2e-07,
      "loss": 0.0162,
      "step": 1481
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1383196542008645,
      "grad_norm": 0.12841884791851044,
      "learning_rate": 2e-07,
      "loss": 0.0472,
      "step": 1482
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13841298730086507,
      "grad_norm": 0.13268157839775085,
      "learning_rate": 2e-07,
      "loss": 0.0471,
      "step": 1483
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13850632040086566,
      "grad_norm": 0.13661326467990875,
      "learning_rate": 2e-07,
      "loss": 0.0489,
      "step": 1484
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13859965350086625,
      "grad_norm": 0.12214091420173645,
      "learning_rate": 2e-07,
      "loss": 0.0501,
      "step": 1485
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13869298660086682,
      "grad_norm": 0.13982731103897095,
      "learning_rate": 2e-07,
      "loss": -0.046,
      "step": 1486
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1387863197008674,
      "grad_norm": 0.11978689581155777,
      "learning_rate": 2e-07,
      "loss": 0.0229,
      "step": 1487
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.138879652800868,
      "grad_norm": 0.13548658788204193,
      "learning_rate": 2e-07,
      "loss": 0.0752,
      "step": 1488
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13897298590086857,
      "grad_norm": 0.1313246190547943,
      "learning_rate": 2e-07,
      "loss": 0.0415,
      "step": 1489
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13906631900086916,
      "grad_norm": 0.12334547936916351,
      "learning_rate": 2e-07,
      "loss": 0.0192,
      "step": 1490
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13915965210086975,
      "grad_norm": 0.1284998506307602,
      "learning_rate": 2e-07,
      "loss": 0.0127,
      "step": 1491
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13925298520087034,
      "grad_norm": 0.132889062166214,
      "learning_rate": 2e-07,
      "loss": 0.008,
      "step": 1492
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1393463183008709,
      "grad_norm": 0.13802756369113922,
      "learning_rate": 2e-07,
      "loss": 0.0654,
      "step": 1493
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1394396514008715,
      "grad_norm": 0.1258484125137329,
      "learning_rate": 2e-07,
      "loss": 0.0397,
      "step": 1494
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1395329845008721,
      "grad_norm": 0.14119204878807068,
      "learning_rate": 2e-07,
      "loss": 0.0943,
      "step": 1495
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13962631760087266,
      "grad_norm": 0.14203029870986938,
      "learning_rate": 2e-07,
      "loss": 0.0544,
      "step": 1496
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13971965070087325,
      "grad_norm": 0.13593608140945435,
      "learning_rate": 2e-07,
      "loss": 0.0188,
      "step": 1497
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.13981298380087384,
      "grad_norm": 0.13752004504203796,
      "learning_rate": 2e-07,
      "loss": -0.0276,
      "step": 1498
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1399063169008744,
      "grad_norm": 0.13230733573436737,
      "learning_rate": 2e-07,
      "loss": -0.0003,
      "step": 1499
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.139999650000875,
      "grad_norm": 0.12464489787817001,
      "learning_rate": 2e-07,
      "loss": 0.0332,
      "step": 1500
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1400929831008756,
      "grad_norm": 0.14144942164421082,
      "learning_rate": 2e-07,
      "loss": 0.0688,
      "step": 1501
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14018631620087615,
      "grad_norm": 0.1382383406162262,
      "learning_rate": 2e-07,
      "loss": 0.0209,
      "step": 1502
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14027964930087675,
      "grad_norm": 0.13249830901622772,
      "learning_rate": 2e-07,
      "loss": 0.027,
      "step": 1503
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14037298240087734,
      "grad_norm": 0.1336497962474823,
      "learning_rate": 2e-07,
      "loss": 0.0216,
      "step": 1504
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.012712751116071397,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4096.0,
      "completions/mean_length": 611.5469970703125,
      "completions/mean_terminated_length": 566.6795654296875,
      "completions/min_length": 91.0,
      "completions/min_terminated_length": 91.0,
      "epoch": 0.1404663155008779,
      "grad_norm": 0.13943322002887726,
      "learning_rate": 2e-07,
      "loss": 0.0564,
      "num_tokens": 1022796423.0,
      "reward": 0.5952671766281128,
      "reward_std": 0.183739572763443,
      "rewards/simpleverify_reward/mean": 0.5952671766281128,
      "rewards/simpleverify_reward/std": 0.4908445477485657,
      "step": 1505
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1405596486008785,
      "grad_norm": 0.12328717112541199,
      "learning_rate": 2e-07,
      "loss": -0.0193,
      "step": 1506
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1406529817008791,
      "grad_norm": 0.13059201836585999,
      "learning_rate": 2e-07,
      "loss": -0.0285,
      "step": 1507
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14074631480087965,
      "grad_norm": 0.12330266833305359,
      "learning_rate": 2e-07,
      "loss": -0.0017,
      "step": 1508
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14083964790088024,
      "grad_norm": 0.13692261278629303,
      "learning_rate": 2e-07,
      "loss": 0.0583,
      "step": 1509
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14093298100088084,
      "grad_norm": 0.14435957372188568,
      "learning_rate": 2e-07,
      "loss": 0.0028,
      "step": 1510
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14102631410088143,
      "grad_norm": 0.11849499493837357,
      "learning_rate": 2e-07,
      "loss": 0.0233,
      "step": 1511
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.141119647200882,
      "grad_norm": 0.1235920712351799,
      "learning_rate": 2e-07,
      "loss": 0.0416,
      "step": 1512
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14121298030088258,
      "grad_norm": 0.12330605834722519,
      "learning_rate": 2e-07,
      "loss": -0.0135,
      "step": 1513
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14130631340088318,
      "grad_norm": 0.12435866892337799,
      "learning_rate": 2e-07,
      "loss": -0.0175,
      "step": 1514
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14139964650088374,
      "grad_norm": 0.1300330013036728,
      "learning_rate": 2e-07,
      "loss": 0.0019,
      "step": 1515
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14149297960088433,
      "grad_norm": 0.1270255744457245,
      "learning_rate": 2e-07,
      "loss": 0.0404,
      "step": 1516
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14158631270088493,
      "grad_norm": 0.13278484344482422,
      "learning_rate": 2e-07,
      "loss": 0.055,
      "step": 1517
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1416796458008855,
      "grad_norm": 0.1299923062324524,
      "learning_rate": 2e-07,
      "loss": -0.0178,
      "step": 1518
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14177297890088608,
      "grad_norm": 0.13562241196632385,
      "learning_rate": 2e-07,
      "loss": 0.0229,
      "step": 1519
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14186631200088667,
      "grad_norm": 0.12498418241739273,
      "learning_rate": 2e-07,
      "loss": 0.0491,
      "step": 1520
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14195964510088724,
      "grad_norm": 0.13515660166740417,
      "learning_rate": 2e-07,
      "loss": 0.038,
      "step": 1521
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14205297820088783,
      "grad_norm": 0.1273382306098938,
      "learning_rate": 2e-07,
      "loss": 0.0904,
      "step": 1522
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14214631130088842,
      "grad_norm": 0.12775585055351257,
      "learning_rate": 2e-07,
      "loss": 0.0804,
      "step": 1523
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.142239644400889,
      "grad_norm": 0.13200220465660095,
      "learning_rate": 2e-07,
      "loss": 0.0376,
      "step": 1524
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14233297750088958,
      "grad_norm": 0.12758012115955353,
      "learning_rate": 2e-07,
      "loss": 0.0164,
      "step": 1525
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14242631060089017,
      "grad_norm": 0.12444531917572021,
      "learning_rate": 2e-07,
      "loss": 0.0083,
      "step": 1526
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14251964370089074,
      "grad_norm": 0.1462988406419754,
      "learning_rate": 2e-07,
      "loss": 0.0337,
      "step": 1527
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14261297680089133,
      "grad_norm": 0.12262451648712158,
      "learning_rate": 2e-07,
      "loss": 0.0468,
      "step": 1528
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14270630990089192,
      "grad_norm": 0.13249804079532623,
      "learning_rate": 2e-07,
      "loss": 0.0675,
      "step": 1529
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14279964300089248,
      "grad_norm": 0.12518811225891113,
      "learning_rate": 2e-07,
      "loss": 0.0297,
      "step": 1530
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14289297610089308,
      "grad_norm": 0.1315504014492035,
      "learning_rate": 2e-07,
      "loss": 0.0685,
      "step": 1531
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14298630920089367,
      "grad_norm": 0.13933368027210236,
      "learning_rate": 2e-07,
      "loss": -0.0068,
      "step": 1532
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14307964230089426,
      "grad_norm": 0.1306922286748886,
      "learning_rate": 2e-07,
      "loss": 0.0372,
      "step": 1533
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14317297540089483,
      "grad_norm": 0.12907284498214722,
      "learning_rate": 2e-07,
      "loss": 0.0044,
      "step": 1534
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14326630850089542,
      "grad_norm": 0.11949587613344193,
      "learning_rate": 2e-07,
      "loss": 0.0582,
      "step": 1535
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.143359641600896,
      "grad_norm": 0.11586958169937134,
      "learning_rate": 2e-07,
      "loss": 0.0154,
      "step": 1536
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14345297470089657,
      "grad_norm": 0.141885906457901,
      "learning_rate": 2e-07,
      "loss": 0.0884,
      "step": 1537
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14354630780089717,
      "grad_norm": 0.13956746459007263,
      "learning_rate": 2e-07,
      "loss": 0.0612,
      "step": 1538
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14363964090089776,
      "grad_norm": 0.13310831785202026,
      "learning_rate": 2e-07,
      "loss": 0.0481,
      "step": 1539
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14373297400089832,
      "grad_norm": 0.12801481783390045,
      "learning_rate": 2e-07,
      "loss": 0.0222,
      "step": 1540
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14382630710089891,
      "grad_norm": 0.1253482550382614,
      "learning_rate": 2e-07,
      "loss": 0.0084,
      "step": 1541
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1439196402008995,
      "grad_norm": 0.13254080712795258,
      "learning_rate": 2e-07,
      "loss": 0.0194,
      "step": 1542
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14401297330090007,
      "grad_norm": 0.12622767686843872,
      "learning_rate": 2e-07,
      "loss": 0.0332,
      "step": 1543
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14410630640090066,
      "grad_norm": 0.131544828414917,
      "learning_rate": 2e-07,
      "loss": 0.0271,
      "step": 1544
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14419963950090126,
      "grad_norm": 0.12295348197221756,
      "learning_rate": 2e-07,
      "loss": 0.0163,
      "step": 1545
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14429297260090182,
      "grad_norm": 0.12304048985242844,
      "learning_rate": 2e-07,
      "loss": 0.0437,
      "step": 1546
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1443863057009024,
      "grad_norm": 0.12579649686813354,
      "learning_rate": 2e-07,
      "loss": 0.0166,
      "step": 1547
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.144479638800903,
      "grad_norm": 0.11576581746339798,
      "learning_rate": 2e-07,
      "loss": -0.0036,
      "step": 1548
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14457297190090357,
      "grad_norm": 0.13440829515457153,
      "learning_rate": 2e-07,
      "loss": 0.0583,
      "step": 1549
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14466630500090416,
      "grad_norm": 0.1419072151184082,
      "learning_rate": 2e-07,
      "loss": 0.0432,
      "step": 1550
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14475963810090475,
      "grad_norm": 0.1361137479543686,
      "learning_rate": 2e-07,
      "loss": 0.0056,
      "step": 1551
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14485297120090532,
      "grad_norm": 0.13247188925743103,
      "learning_rate": 2e-07,
      "loss": 0.0514,
      "step": 1552
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1449463043009059,
      "grad_norm": 0.12846198678016663,
      "learning_rate": 2e-07,
      "loss": 0.0092,
      "step": 1553
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1450396374009065,
      "grad_norm": 0.11642246693372726,
      "learning_rate": 2e-07,
      "loss": 0.072,
      "step": 1554
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1451329705009071,
      "grad_norm": 0.12781000137329102,
      "learning_rate": 2e-07,
      "loss": -0.0145,
      "step": 1555
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14522630360090766,
      "grad_norm": 0.13026975095272064,
      "learning_rate": 2e-07,
      "loss": 0.0429,
      "step": 1556
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14531963670090825,
      "grad_norm": 0.1336844265460968,
      "learning_rate": 2e-07,
      "loss": -0.0242,
      "step": 1557
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14541296980090884,
      "grad_norm": 0.13330115377902985,
      "learning_rate": 2e-07,
      "loss": 0.0334,
      "step": 1558
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1455063029009094,
      "grad_norm": 0.13167071342468262,
      "learning_rate": 2e-07,
      "loss": 0.014,
      "step": 1559
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14559963600091,
      "grad_norm": 0.1251688450574875,
      "learning_rate": 2e-07,
      "loss": 0.0574,
      "step": 1560
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1456929691009106,
      "grad_norm": 0.12467975169420242,
      "learning_rate": 2e-07,
      "loss": 0.02,
      "step": 1561
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14578630220091116,
      "grad_norm": 0.1277707815170288,
      "learning_rate": 2e-07,
      "loss": 0.0408,
      "step": 1562
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14587963530091175,
      "grad_norm": 0.1245904341340065,
      "learning_rate": 2e-07,
      "loss": 0.006,
      "step": 1563
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14597296840091234,
      "grad_norm": 0.14062905311584473,
      "learning_rate": 2e-07,
      "loss": 0.014,
      "step": 1564
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1460663015009129,
      "grad_norm": 0.1285715103149414,
      "learning_rate": 2e-07,
      "loss": -0.0086,
      "step": 1565
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1461596346009135,
      "grad_norm": 0.1425938755273819,
      "learning_rate": 2e-07,
      "loss": -0.0099,
      "step": 1566
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1462529677009141,
      "grad_norm": 0.12996259331703186,
      "learning_rate": 2e-07,
      "loss": 0.0447,
      "step": 1567
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14634630080091465,
      "grad_norm": 0.12410619109869003,
      "learning_rate": 2e-07,
      "loss": -0.0095,
      "step": 1568
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.012102399553571397,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4087.0,
      "completions/mean_length": 614.2689208984375,
      "completions/mean_terminated_length": 571.6153564453125,
      "completions/min_length": 84.0,
      "completions/min_terminated_length": 84.0,
      "epoch": 0.14643963390091524,
      "grad_norm": 0.1255834847688675,
      "learning_rate": 2e-07,
      "loss": 0.0413,
      "num_tokens": 1063670202.0,
      "reward": 0.5979527235031128,
      "reward_std": 0.1826418787240982,
      "rewards/simpleverify_reward/mean": 0.5979527235031128,
      "rewards/simpleverify_reward/std": 0.490315705537796,
      "step": 1569
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14653296700091584,
      "grad_norm": 0.1274181604385376,
      "learning_rate": 2e-07,
      "loss": 0.0088,
      "step": 1570
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1466263001009164,
      "grad_norm": 0.126582533121109,
      "learning_rate": 2e-07,
      "loss": 0.0316,
      "step": 1571
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.146719633200917,
      "grad_norm": 0.13323159515857697,
      "learning_rate": 2e-07,
      "loss": 0.0302,
      "step": 1572
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14681296630091759,
      "grad_norm": 0.12743723392486572,
      "learning_rate": 2e-07,
      "loss": 0.0898,
      "step": 1573
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14690629940091818,
      "grad_norm": 0.12763725221157074,
      "learning_rate": 2e-07,
      "loss": 0.0363,
      "step": 1574
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14699963250091874,
      "grad_norm": 0.11844750493764877,
      "learning_rate": 2e-07,
      "loss": 0.0203,
      "step": 1575
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14709296560091933,
      "grad_norm": 0.12216989696025848,
      "learning_rate": 2e-07,
      "loss": 0.0264,
      "step": 1576
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14718629870091993,
      "grad_norm": 0.12879298627376556,
      "learning_rate": 2e-07,
      "loss": 0.0503,
      "step": 1577
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1472796318009205,
      "grad_norm": 0.13290274143218994,
      "learning_rate": 2e-07,
      "loss": 0.0776,
      "step": 1578
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14737296490092108,
      "grad_norm": 0.13088519871234894,
      "learning_rate": 2e-07,
      "loss": 0.0466,
      "step": 1579
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14746629800092168,
      "grad_norm": 0.130288764834404,
      "learning_rate": 2e-07,
      "loss": 0.0197,
      "step": 1580
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14755963110092224,
      "grad_norm": 0.1316644251346588,
      "learning_rate": 2e-07,
      "loss": 0.0481,
      "step": 1581
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14765296420092283,
      "grad_norm": 0.12390497326850891,
      "learning_rate": 2e-07,
      "loss": 0.0011,
      "step": 1582
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14774629730092342,
      "grad_norm": 0.13405756652355194,
      "learning_rate": 2e-07,
      "loss": 0.0033,
      "step": 1583
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.147839630400924,
      "grad_norm": 0.13446617126464844,
      "learning_rate": 2e-07,
      "loss": -0.004,
      "step": 1584
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14793296350092458,
      "grad_norm": 0.1343328207731247,
      "learning_rate": 2e-07,
      "loss": 0.0169,
      "step": 1585
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14802629660092517,
      "grad_norm": 0.16320449113845825,
      "learning_rate": 2e-07,
      "loss": 0.0245,
      "step": 1586
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14811962970092574,
      "grad_norm": 0.12486258149147034,
      "learning_rate": 2e-07,
      "loss": 0.0286,
      "step": 1587
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14821296280092633,
      "grad_norm": 0.11726725846529007,
      "learning_rate": 2e-07,
      "loss": 0.0086,
      "step": 1588
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14830629590092692,
      "grad_norm": 0.1314283311367035,
      "learning_rate": 2e-07,
      "loss": 0.0489,
      "step": 1589
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14839962900092749,
      "grad_norm": 0.1410963237285614,
      "learning_rate": 2e-07,
      "loss": 0.0157,
      "step": 1590
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14849296210092808,
      "grad_norm": 0.1482408493757248,
      "learning_rate": 2e-07,
      "loss": 0.0194,
      "step": 1591
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14858629520092867,
      "grad_norm": 0.1295221894979477,
      "learning_rate": 2e-07,
      "loss": 0.03,
      "step": 1592
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14867962830092923,
      "grad_norm": 0.12675008177757263,
      "learning_rate": 2e-07,
      "loss": 0.0152,
      "step": 1593
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14877296140092983,
      "grad_norm": 0.1261344999074936,
      "learning_rate": 2e-07,
      "loss": 0.0292,
      "step": 1594
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14886629450093042,
      "grad_norm": 0.11228286474943161,
      "learning_rate": 2e-07,
      "loss": 0.0003,
      "step": 1595
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.148959627600931,
      "grad_norm": 0.13561365008354187,
      "learning_rate": 2e-07,
      "loss": -0.0352,
      "step": 1596
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14905296070093157,
      "grad_norm": 0.11695383489131927,
      "learning_rate": 2e-07,
      "loss": 0.0326,
      "step": 1597
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14914629380093217,
      "grad_norm": 0.1335514932870865,
      "learning_rate": 2e-07,
      "loss": 0.0571,
      "step": 1598
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14923962690093276,
      "grad_norm": 0.1267106533050537,
      "learning_rate": 2e-07,
      "loss": 0.0173,
      "step": 1599
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14933296000093332,
      "grad_norm": 0.1275610476732254,
      "learning_rate": 2e-07,
      "loss": 0.0153,
      "step": 1600
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14942629310093392,
      "grad_norm": 0.15320239961147308,
      "learning_rate": 2e-07,
      "loss": 0.0447,
      "step": 1601
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1495196262009345,
      "grad_norm": 0.13889829814434052,
      "learning_rate": 2e-07,
      "loss": 0.0228,
      "step": 1602
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14961295930093507,
      "grad_norm": 0.14359885454177856,
      "learning_rate": 2e-07,
      "loss": 0.0692,
      "step": 1603
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14970629240093566,
      "grad_norm": 0.12624703347682953,
      "learning_rate": 2e-07,
      "loss": 0.0049,
      "step": 1604
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14979962550093626,
      "grad_norm": 0.1442633420228958,
      "learning_rate": 2e-07,
      "loss": 0.0469,
      "step": 1605
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.14989295860093682,
      "grad_norm": 0.1283549815416336,
      "learning_rate": 2e-07,
      "loss": 0.0172,
      "step": 1606
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1499862917009374,
      "grad_norm": 0.12885038554668427,
      "learning_rate": 2e-07,
      "loss": 0.0271,
      "step": 1607
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.150079624800938,
      "grad_norm": 0.1230170726776123,
      "learning_rate": 2e-07,
      "loss": 0.0375,
      "step": 1608
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15017295790093857,
      "grad_norm": 0.1217385083436966,
      "learning_rate": 2e-07,
      "loss": 0.0408,
      "step": 1609
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15026629100093916,
      "grad_norm": 0.1265076845884323,
      "learning_rate": 2e-07,
      "loss": -0.0227,
      "step": 1610
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15035962410093975,
      "grad_norm": 0.13014468550682068,
      "learning_rate": 2e-07,
      "loss": 0.0351,
      "step": 1611
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15045295720094032,
      "grad_norm": 0.13369403779506683,
      "learning_rate": 2e-07,
      "loss": 0.0302,
      "step": 1612
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1505462903009409,
      "grad_norm": 0.12770317494869232,
      "learning_rate": 2e-07,
      "loss": -0.0355,
      "step": 1613
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1506396234009415,
      "grad_norm": 0.13406477868556976,
      "learning_rate": 2e-07,
      "loss": 0.011,
      "step": 1614
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1507329565009421,
      "grad_norm": 0.1345720887184143,
      "learning_rate": 2e-07,
      "loss": 0.0324,
      "step": 1615
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15082628960094266,
      "grad_norm": 0.13447582721710205,
      "learning_rate": 2e-07,
      "loss": 0.058,
      "step": 1616
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15091962270094325,
      "grad_norm": 0.14136497676372528,
      "learning_rate": 2e-07,
      "loss": 0.0127,
      "step": 1617
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15101295580094384,
      "grad_norm": 0.13039113581180573,
      "learning_rate": 2e-07,
      "loss": 0.0265,
      "step": 1618
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1511062889009444,
      "grad_norm": 0.13440050184726715,
      "learning_rate": 2e-07,
      "loss": 0.0415,
      "step": 1619
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.151199622000945,
      "grad_norm": 0.1323552429676056,
      "learning_rate": 2e-07,
      "loss": 0.027,
      "step": 1620
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1512929551009456,
      "grad_norm": 0.144898921251297,
      "learning_rate": 2e-07,
      "loss": 0.0341,
      "step": 1621
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15138628820094616,
      "grad_norm": 0.14660273492336273,
      "learning_rate": 2e-07,
      "loss": -0.0229,
      "step": 1622
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15147962130094675,
      "grad_norm": 0.1274280548095703,
      "learning_rate": 2e-07,
      "loss": 0.0251,
      "step": 1623
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15157295440094734,
      "grad_norm": 0.12751837074756622,
      "learning_rate": 2e-07,
      "loss": -0.0064,
      "step": 1624
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1516662875009479,
      "grad_norm": 0.12840381264686584,
      "learning_rate": 2e-07,
      "loss": 0.0723,
      "step": 1625
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1517596206009485,
      "grad_norm": 0.13688528537750244,
      "learning_rate": 2e-07,
      "loss": 0.0275,
      "step": 1626
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1518529537009491,
      "grad_norm": 0.1506073921918869,
      "learning_rate": 2e-07,
      "loss": 0.0403,
      "step": 1627
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15194628680094965,
      "grad_norm": 0.1503155380487442,
      "learning_rate": 2e-07,
      "loss": 0.0364,
      "step": 1628
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15203961990095025,
      "grad_norm": 0.13525450229644775,
      "learning_rate": 2e-07,
      "loss": 0.0368,
      "step": 1629
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15213295300095084,
      "grad_norm": 0.14401701092720032,
      "learning_rate": 2e-07,
      "loss": 0.0278,
      "step": 1630
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1522262861009514,
      "grad_norm": 0.1395573616027832,
      "learning_rate": 2e-07,
      "loss": -0.0194,
      "step": 1631
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.152319619200952,
      "grad_norm": 0.14466142654418945,
      "learning_rate": 2e-07,
      "loss": 0.0168,
      "step": 1632
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.011300223214285698,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4088.0,
      "completions/mean_length": 607.0234375,
      "completions/mean_terminated_length": 567.1466674804688,
      "completions/min_length": 85.0,
      "completions/min_terminated_length": 85.0,
      "epoch": 0.1524129523009526,
      "grad_norm": 0.12688760459423065,
      "learning_rate": 2e-07,
      "loss": 0.0459,
      "num_tokens": 1104186637.0,
      "reward": 0.6036028265953064,
      "reward_std": 0.18167415261268616,
      "rewards/simpleverify_reward/mean": 0.6036028265953064,
      "rewards/simpleverify_reward/std": 0.4891529977321625,
      "step": 1633
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15250628540095315,
      "grad_norm": 0.1347307711839676,
      "learning_rate": 2e-07,
      "loss": 0.0385,
      "step": 1634
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15259961850095374,
      "grad_norm": 0.1331218034029007,
      "learning_rate": 2e-07,
      "loss": 0.0188,
      "step": 1635
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15269295160095434,
      "grad_norm": 0.12972086668014526,
      "learning_rate": 2e-07,
      "loss": -0.0072,
      "step": 1636
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15278628470095493,
      "grad_norm": 0.12005738168954849,
      "learning_rate": 2e-07,
      "loss": 0.0089,
      "step": 1637
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1528796178009555,
      "grad_norm": 0.1425633728504181,
      "learning_rate": 2e-07,
      "loss": 0.0432,
      "step": 1638
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15297295090095608,
      "grad_norm": 0.1493898630142212,
      "learning_rate": 2e-07,
      "loss": 0.045,
      "step": 1639
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15306628400095668,
      "grad_norm": 0.1384979635477066,
      "learning_rate": 2e-07,
      "loss": -0.0196,
      "step": 1640
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15315961710095724,
      "grad_norm": 0.12740084528923035,
      "learning_rate": 2e-07,
      "loss": 0.0237,
      "step": 1641
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15325295020095783,
      "grad_norm": 0.13496433198451996,
      "learning_rate": 2e-07,
      "loss": 0.0684,
      "step": 1642
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15334628330095842,
      "grad_norm": 0.12255571037530899,
      "learning_rate": 2e-07,
      "loss": 0.0213,
      "step": 1643
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.153439616400959,
      "grad_norm": 0.12416941672563553,
      "learning_rate": 2e-07,
      "loss": 0.0114,
      "step": 1644
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15353294950095958,
      "grad_norm": 0.12521426379680634,
      "learning_rate": 2e-07,
      "loss": 0.0226,
      "step": 1645
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15362628260096017,
      "grad_norm": 0.12777189910411835,
      "learning_rate": 2e-07,
      "loss": 0.0468,
      "step": 1646
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15371961570096074,
      "grad_norm": 0.1303749531507492,
      "learning_rate": 2e-07,
      "loss": 0.0167,
      "step": 1647
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15381294880096133,
      "grad_norm": 0.12536653876304626,
      "learning_rate": 2e-07,
      "loss": 0.0759,
      "step": 1648
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15390628190096192,
      "grad_norm": 0.14741076529026031,
      "learning_rate": 2e-07,
      "loss": 0.0454,
      "step": 1649
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1539996150009625,
      "grad_norm": 0.12200335413217545,
      "learning_rate": 2e-07,
      "loss": 0.0044,
      "step": 1650
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15409294810096308,
      "grad_norm": 0.13597562909126282,
      "learning_rate": 2e-07,
      "loss": 0.0445,
      "step": 1651
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15418628120096367,
      "grad_norm": 0.1319851577281952,
      "learning_rate": 2e-07,
      "loss": 0.0355,
      "step": 1652
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15427961430096424,
      "grad_norm": 0.15145090222358704,
      "learning_rate": 2e-07,
      "loss": 0.0221,
      "step": 1653
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15437294740096483,
      "grad_norm": 0.12274263799190521,
      "learning_rate": 2e-07,
      "loss": 0.0358,
      "step": 1654
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15446628050096542,
      "grad_norm": 0.13365498185157776,
      "learning_rate": 2e-07,
      "loss": 0.0548,
      "step": 1655
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15455961360096598,
      "grad_norm": 0.13028521835803986,
      "learning_rate": 2e-07,
      "loss": 0.0694,
      "step": 1656
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15465294670096658,
      "grad_norm": 0.13391505181789398,
      "learning_rate": 2e-07,
      "loss": 0.036,
      "step": 1657
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15474627980096717,
      "grad_norm": 0.12943685054779053,
      "learning_rate": 2e-07,
      "loss": 0.0018,
      "step": 1658
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15483961290096776,
      "grad_norm": 0.1325240135192871,
      "learning_rate": 2e-07,
      "loss": 0.0232,
      "step": 1659
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15493294600096832,
      "grad_norm": 0.12339305132627487,
      "learning_rate": 2e-07,
      "loss": 0.0233,
      "step": 1660
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15502627910096892,
      "grad_norm": 0.12262265384197235,
      "learning_rate": 2e-07,
      "loss": 0.0404,
      "step": 1661
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1551196122009695,
      "grad_norm": 0.13826113939285278,
      "learning_rate": 2e-07,
      "loss": 0.0393,
      "step": 1662
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15521294530097007,
      "grad_norm": 0.1265203207731247,
      "learning_rate": 2e-07,
      "loss": 0.0413,
      "step": 1663
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15530627840097067,
      "grad_norm": 0.12628337740898132,
      "learning_rate": 2e-07,
      "loss": -0.0049,
      "step": 1664
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15539961150097126,
      "grad_norm": 0.13126251101493835,
      "learning_rate": 2e-07,
      "loss": 0.0898,
      "step": 1665
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15549294460097182,
      "grad_norm": 0.12042747437953949,
      "learning_rate": 2e-07,
      "loss": -0.0074,
      "step": 1666
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15558627770097241,
      "grad_norm": 0.1237117350101471,
      "learning_rate": 2e-07,
      "loss": 0.0042,
      "step": 1667
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.155679610800973,
      "grad_norm": 0.1371576488018036,
      "learning_rate": 2e-07,
      "loss": 0.0465,
      "step": 1668
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15577294390097357,
      "grad_norm": 0.14915134012699127,
      "learning_rate": 2e-07,
      "loss": 0.0424,
      "step": 1669
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15586627700097416,
      "grad_norm": 0.11963173002004623,
      "learning_rate": 2e-07,
      "loss": 0.0036,
      "step": 1670
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15595961010097475,
      "grad_norm": 0.13197095692157745,
      "learning_rate": 2e-07,
      "loss": 0.061,
      "step": 1671
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15605294320097532,
      "grad_norm": 0.12244408577680588,
      "learning_rate": 2e-07,
      "loss": 0.0374,
      "step": 1672
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1561462763009759,
      "grad_norm": 0.11905831098556519,
      "learning_rate": 2e-07,
      "loss": 0.0247,
      "step": 1673
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1562396094009765,
      "grad_norm": 0.1333603411912918,
      "learning_rate": 2e-07,
      "loss": 0.0248,
      "step": 1674
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15633294250097707,
      "grad_norm": 0.1352948248386383,
      "learning_rate": 2e-07,
      "loss": 0.0265,
      "step": 1675
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15642627560097766,
      "grad_norm": 0.12792669236660004,
      "learning_rate": 2e-07,
      "loss": 0.0017,
      "step": 1676
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15651960870097825,
      "grad_norm": 0.12432333827018738,
      "learning_rate": 2e-07,
      "loss": 0.0217,
      "step": 1677
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15661294180097884,
      "grad_norm": 0.13010576367378235,
      "learning_rate": 2e-07,
      "loss": 0.0282,
      "step": 1678
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1567062749009794,
      "grad_norm": 0.12494930624961853,
      "learning_rate": 2e-07,
      "loss": -0.0043,
      "step": 1679
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15679960800098,
      "grad_norm": 0.13553807139396667,
      "learning_rate": 2e-07,
      "loss": 0.0217,
      "step": 1680
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1568929411009806,
      "grad_norm": 0.13172224164009094,
      "learning_rate": 2e-07,
      "loss": 0.0329,
      "step": 1681
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15698627420098116,
      "grad_norm": 0.13221698999404907,
      "learning_rate": 2e-07,
      "loss": 0.0364,
      "step": 1682
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15707960730098175,
      "grad_norm": 0.1474255472421646,
      "learning_rate": 2e-07,
      "loss": 0.0274,
      "step": 1683
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15717294040098234,
      "grad_norm": 0.12861399352550507,
      "learning_rate": 2e-07,
      "loss": 0.0088,
      "step": 1684
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1572662735009829,
      "grad_norm": 0.15298102796077728,
      "learning_rate": 2e-07,
      "loss": 0.0064,
      "step": 1685
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1573596066009835,
      "grad_norm": 0.13587716221809387,
      "learning_rate": 2e-07,
      "loss": -0.0259,
      "step": 1686
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1574529397009841,
      "grad_norm": 0.1306609809398651,
      "learning_rate": 2e-07,
      "loss": 0.0436,
      "step": 1687
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15754627280098465,
      "grad_norm": 0.13562940061092377,
      "learning_rate": 2e-07,
      "loss": 0.03,
      "step": 1688
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15763960590098525,
      "grad_norm": 0.12904827296733856,
      "learning_rate": 2e-07,
      "loss": -0.0048,
      "step": 1689
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15773293900098584,
      "grad_norm": 0.13245359063148499,
      "learning_rate": 2e-07,
      "loss": -0.0109,
      "step": 1690
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1578262721009864,
      "grad_norm": 0.12895280122756958,
      "learning_rate": 2e-07,
      "loss": 0.009,
      "step": 1691
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.157919605200987,
      "grad_norm": 0.128696471452713,
      "learning_rate": 2e-07,
      "loss": 0.0514,
      "step": 1692
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1580129383009876,
      "grad_norm": 0.26684674620628357,
      "learning_rate": 2e-07,
      "loss": 0.0167,
      "step": 1693
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15810627140098815,
      "grad_norm": 0.13796038925647736,
      "learning_rate": 2e-07,
      "loss": 0.0098,
      "step": 1694
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15819960450098874,
      "grad_norm": 0.11801528930664062,
      "learning_rate": 2e-07,
      "loss": 0.0306,
      "step": 1695
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15829293760098934,
      "grad_norm": 0.13526540994644165,
      "learning_rate": 2e-07,
      "loss": -0.0183,
      "step": 1696
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.011579241071428603,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4089.0,
      "completions/mean_length": 612.7501220703125,
      "completions/mean_terminated_length": 571.9441528320312,
      "completions/min_length": 76.0,
      "completions/min_terminated_length": 76.0,
      "epoch": 0.1583862707009899,
      "grad_norm": 0.1290045976638794,
      "learning_rate": 2e-07,
      "loss": 0.0604,
      "num_tokens": 1145024370.0,
      "reward": 0.600045382976532,
      "reward_std": 0.18511967360973358,
      "rewards/simpleverify_reward/mean": 0.6000453233718872,
      "rewards/simpleverify_reward/std": 0.48989295959472656,
      "step": 1697
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1584796038009905,
      "grad_norm": 0.12889961898326874,
      "learning_rate": 2e-07,
      "loss": 0.0152,
      "step": 1698
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15857293690099108,
      "grad_norm": 0.14831697940826416,
      "learning_rate": 2e-07,
      "loss": -0.031,
      "step": 1699
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15866627000099168,
      "grad_norm": 0.12529629468917847,
      "learning_rate": 2e-07,
      "loss": -0.0083,
      "step": 1700
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15875960310099224,
      "grad_norm": 0.12842491269111633,
      "learning_rate": 2e-07,
      "loss": 0.0244,
      "step": 1701
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15885293620099283,
      "grad_norm": 0.12892888486385345,
      "learning_rate": 2e-07,
      "loss": 0.0271,
      "step": 1702
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15894626930099343,
      "grad_norm": 0.13366591930389404,
      "learning_rate": 2e-07,
      "loss": -0.0035,
      "step": 1703
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.159039602400994,
      "grad_norm": 0.12670020759105682,
      "learning_rate": 2e-07,
      "loss": 0.01,
      "step": 1704
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15913293550099458,
      "grad_norm": 0.125052809715271,
      "learning_rate": 2e-07,
      "loss": 0.0182,
      "step": 1705
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15922626860099517,
      "grad_norm": 0.15874795615673065,
      "learning_rate": 2e-07,
      "loss": 0.0309,
      "step": 1706
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15931960170099574,
      "grad_norm": 0.1257004290819168,
      "learning_rate": 2e-07,
      "loss": 0.0212,
      "step": 1707
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15941293480099633,
      "grad_norm": 0.12522895634174347,
      "learning_rate": 2e-07,
      "loss": 0.0208,
      "step": 1708
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15950626790099692,
      "grad_norm": 0.14015363156795502,
      "learning_rate": 2e-07,
      "loss": 0.0433,
      "step": 1709
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1595996010009975,
      "grad_norm": 0.12972092628479004,
      "learning_rate": 2e-07,
      "loss": -0.0161,
      "step": 1710
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15969293410099808,
      "grad_norm": 0.13276153802871704,
      "learning_rate": 2e-07,
      "loss": 0.0178,
      "step": 1711
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15978626720099867,
      "grad_norm": 0.13770943880081177,
      "learning_rate": 2e-07,
      "loss": 0.0452,
      "step": 1712
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15987960030099924,
      "grad_norm": 0.12955541908740997,
      "learning_rate": 2e-07,
      "loss": 0.0334,
      "step": 1713
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.15997293340099983,
      "grad_norm": 0.12294356524944305,
      "learning_rate": 2e-07,
      "loss": 0.0232,
      "step": 1714
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16006626650100042,
      "grad_norm": 0.12066502869129181,
      "learning_rate": 2e-07,
      "loss": 0.0322,
      "step": 1715
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16015959960100098,
      "grad_norm": 0.1262289434671402,
      "learning_rate": 2e-07,
      "loss": -0.0043,
      "step": 1716
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16025293270100158,
      "grad_norm": 0.14107133448123932,
      "learning_rate": 2e-07,
      "loss": 0.0102,
      "step": 1717
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16034626580100217,
      "grad_norm": 0.14218828082084656,
      "learning_rate": 2e-07,
      "loss": 0.0111,
      "step": 1718
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16043959890100276,
      "grad_norm": 0.14309479296207428,
      "learning_rate": 2e-07,
      "loss": 0.0131,
      "step": 1719
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16053293200100333,
      "grad_norm": 0.1403043270111084,
      "learning_rate": 2e-07,
      "loss": 0.1061,
      "step": 1720
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16062626510100392,
      "grad_norm": 0.13049417734146118,
      "learning_rate": 2e-07,
      "loss": 0.0245,
      "step": 1721
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1607195982010045,
      "grad_norm": 0.1225542277097702,
      "learning_rate": 2e-07,
      "loss": 0.0147,
      "step": 1722
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16081293130100507,
      "grad_norm": 0.13677062094211578,
      "learning_rate": 2e-07,
      "loss": 0.0073,
      "step": 1723
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16090626440100567,
      "grad_norm": 0.1378423124551773,
      "learning_rate": 2e-07,
      "loss": 0.0263,
      "step": 1724
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16099959750100626,
      "grad_norm": 0.13612021505832672,
      "learning_rate": 2e-07,
      "loss": 0.0208,
      "step": 1725
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16109293060100682,
      "grad_norm": 0.13025879859924316,
      "learning_rate": 2e-07,
      "loss": 0.0352,
      "step": 1726
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16118626370100742,
      "grad_norm": 0.144289031624794,
      "learning_rate": 2e-07,
      "loss": 0.0164,
      "step": 1727
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.161279596801008,
      "grad_norm": 0.1390911191701889,
      "learning_rate": 2e-07,
      "loss": 0.0224,
      "step": 1728
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16137292990100857,
      "grad_norm": 0.12302779406309128,
      "learning_rate": 2e-07,
      "loss": -0.0004,
      "step": 1729
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16146626300100916,
      "grad_norm": 0.1456684023141861,
      "learning_rate": 2e-07,
      "loss": 0.0059,
      "step": 1730
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16155959610100976,
      "grad_norm": 0.1347608119249344,
      "learning_rate": 2e-07,
      "loss": 0.0385,
      "step": 1731
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16165292920101032,
      "grad_norm": 0.14202958345413208,
      "learning_rate": 2e-07,
      "loss": 0.0581,
      "step": 1732
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1617462623010109,
      "grad_norm": 0.13702383637428284,
      "learning_rate": 2e-07,
      "loss": 0.0365,
      "step": 1733
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1618395954010115,
      "grad_norm": 0.13873878121376038,
      "learning_rate": 2e-07,
      "loss": 0.0259,
      "step": 1734
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16193292850101207,
      "grad_norm": 0.14012522995471954,
      "learning_rate": 2e-07,
      "loss": 0.0488,
      "step": 1735
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16202626160101266,
      "grad_norm": 0.12661084532737732,
      "learning_rate": 2e-07,
      "loss": 0.0148,
      "step": 1736
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16211959470101325,
      "grad_norm": 0.13942812383174896,
      "learning_rate": 2e-07,
      "loss": 0.0093,
      "step": 1737
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16221292780101382,
      "grad_norm": 0.1348314881324768,
      "learning_rate": 2e-07,
      "loss": 0.0025,
      "step": 1738
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1623062609010144,
      "grad_norm": 0.14213192462921143,
      "learning_rate": 2e-07,
      "loss": 0.0373,
      "step": 1739
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.162399594001015,
      "grad_norm": 0.1281989961862564,
      "learning_rate": 2e-07,
      "loss": 0.0614,
      "step": 1740
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1624929271010156,
      "grad_norm": 0.15294545888900757,
      "learning_rate": 2e-07,
      "loss": -0.0002,
      "step": 1741
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16258626020101616,
      "grad_norm": 0.1489385962486267,
      "learning_rate": 2e-07,
      "loss": 0.0383,
      "step": 1742
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16267959330101675,
      "grad_norm": 0.14791050553321838,
      "learning_rate": 2e-07,
      "loss": 0.0546,
      "step": 1743
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16277292640101734,
      "grad_norm": 0.13886301219463348,
      "learning_rate": 2e-07,
      "loss": 0.0453,
      "step": 1744
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1628662595010179,
      "grad_norm": 0.16963832080364227,
      "learning_rate": 2e-07,
      "loss": 0.0487,
      "step": 1745
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1629595926010185,
      "grad_norm": 0.1556134670972824,
      "learning_rate": 2e-07,
      "loss": 0.0122,
      "step": 1746
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1630529257010191,
      "grad_norm": 0.14118660986423492,
      "learning_rate": 2e-07,
      "loss": 0.0362,
      "step": 1747
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16314625880101966,
      "grad_norm": 0.13312576711177826,
      "learning_rate": 2e-07,
      "loss": -0.0247,
      "step": 1748
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16323959190102025,
      "grad_norm": 0.13028459250926971,
      "learning_rate": 2e-07,
      "loss": 0.0041,
      "step": 1749
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16333292500102084,
      "grad_norm": 0.1503874510526657,
      "learning_rate": 2e-07,
      "loss": 0.0951,
      "step": 1750
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1634262581010214,
      "grad_norm": 0.1632710099220276,
      "learning_rate": 2e-07,
      "loss": 0.0346,
      "step": 1751
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.163519591201022,
      "grad_norm": 0.1434256136417389,
      "learning_rate": 2e-07,
      "loss": 0.0327,
      "step": 1752
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1636129243010226,
      "grad_norm": 0.1334051638841629,
      "learning_rate": 2e-07,
      "loss": 0.0567,
      "step": 1753
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16370625740102315,
      "grad_norm": 0.13846084475517273,
      "learning_rate": 2e-07,
      "loss": 0.0048,
      "step": 1754
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16379959050102375,
      "grad_norm": 0.14348450303077698,
      "learning_rate": 2e-07,
      "loss": 0.0007,
      "step": 1755
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16389292360102434,
      "grad_norm": 0.13149744272232056,
      "learning_rate": 2e-07,
      "loss": 0.0598,
      "step": 1756
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1639862567010249,
      "grad_norm": 0.14540605247020721,
      "learning_rate": 2e-07,
      "loss": -0.0143,
      "step": 1757
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1640795898010255,
      "grad_norm": 0.15157799422740936,
      "learning_rate": 2e-07,
      "loss": 0.0368,
      "step": 1758
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16417292290102609,
      "grad_norm": 0.15297657251358032,
      "learning_rate": 2e-07,
      "loss": 0.0196,
      "step": 1759
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16426625600102665,
      "grad_norm": 0.1308118999004364,
      "learning_rate": 2e-07,
      "loss": 0.0305,
      "step": 1760
    },
    {
      "clip_ratio/high_max": 0.018076628795824945,
      "clip_ratio/high_mean": 0.008838482841383666,
      "clip_ratio/low_mean": 0.003733100224053487,
      "clip_ratio/low_min": 0.0004693133814726025,
      "clip_ratio/region_mean": 0.012571583036333323,
      "completions/clipped_ratio": 0.011038643973214302,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4089.0,
      "completions/mean_length": 611.9730834960938,
      "completions/mean_terminated_length": 573.0847778320312,
      "completions/min_length": 78.0,
      "completions/min_terminated_length": 78.0,
      "epoch": 0.16435958910102724,
      "grad_norm": 0.10833572596311569,
      "learning_rate": 2e-07,
      "loss": 0.0141,
      "num_tokens": 1185818079.0,
      "reward": 0.6115373969078064,
      "reward_std": 0.180352583527565,
      "rewards/simpleverify_reward/mean": 0.6115373969078064,
      "rewards/simpleverify_reward/std": 0.4874049425125122,
      "step": 1761
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16445292220102783,
      "grad_norm": 0.1127336174249649,
      "learning_rate": 2e-07,
      "loss": 0.0355,
      "step": 1762
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16454625530102843,
      "grad_norm": 0.13127727806568146,
      "learning_rate": 2e-07,
      "loss": 0.0356,
      "step": 1763
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.164639588401029,
      "grad_norm": 0.13594135642051697,
      "learning_rate": 2e-07,
      "loss": 0.0436,
      "step": 1764
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16473292150102958,
      "grad_norm": 0.1298554688692093,
      "learning_rate": 2e-07,
      "loss": -0.0052,
      "step": 1765
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16482625460103018,
      "grad_norm": 0.12493173032999039,
      "learning_rate": 2e-07,
      "loss": 0.0685,
      "step": 1766
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16491958770103074,
      "grad_norm": 0.13796378672122955,
      "learning_rate": 2e-07,
      "loss": -0.0158,
      "step": 1767
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16501292080103133,
      "grad_norm": 0.13567331433296204,
      "learning_rate": 2e-07,
      "loss": 0.0082,
      "step": 1768
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16510625390103192,
      "grad_norm": 0.13345091044902802,
      "learning_rate": 2e-07,
      "loss": 0.0117,
      "step": 1769
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1651995870010325,
      "grad_norm": 0.13981086015701294,
      "learning_rate": 2e-07,
      "loss": 0.0136,
      "step": 1770
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16529292010103308,
      "grad_norm": 0.14303749799728394,
      "learning_rate": 2e-07,
      "loss": 0.0095,
      "step": 1771
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16538625320103367,
      "grad_norm": 0.12364780157804489,
      "learning_rate": 2e-07,
      "loss": 0.028,
      "step": 1772
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16547958630103424,
      "grad_norm": 0.14133331179618835,
      "learning_rate": 2e-07,
      "loss": 0.0188,
      "step": 1773
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16557291940103483,
      "grad_norm": 0.13271479308605194,
      "learning_rate": 2e-07,
      "loss": -0.0079,
      "step": 1774
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16566625250103542,
      "grad_norm": 0.12968213856220245,
      "learning_rate": 2e-07,
      "loss": 0.0744,
      "step": 1775
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16575958560103599,
      "grad_norm": 0.13478386402130127,
      "learning_rate": 2e-07,
      "loss": 0.0292,
      "step": 1776
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16585291870103658,
      "grad_norm": 0.13739782571792603,
      "learning_rate": 2e-07,
      "loss": 0.022,
      "step": 1777
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16594625180103717,
      "grad_norm": 0.15036891400814056,
      "learning_rate": 2e-07,
      "loss": -0.0148,
      "step": 1778
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16603958490103773,
      "grad_norm": 0.12279122322797775,
      "learning_rate": 2e-07,
      "loss": 0.0167,
      "step": 1779
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16613291800103833,
      "grad_norm": 0.14099600911140442,
      "learning_rate": 2e-07,
      "loss": 0.0456,
      "step": 1780
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16622625110103892,
      "grad_norm": 0.17009855806827545,
      "learning_rate": 2e-07,
      "loss": 0.0197,
      "step": 1781
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1663195842010395,
      "grad_norm": 0.13186194002628326,
      "learning_rate": 2e-07,
      "loss": 0.0417,
      "step": 1782
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16641291730104008,
      "grad_norm": 0.12062453478574753,
      "learning_rate": 2e-07,
      "loss": -0.0157,
      "step": 1783
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16650625040104067,
      "grad_norm": 0.1374969184398651,
      "learning_rate": 2e-07,
      "loss": 0.0262,
      "step": 1784
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16659958350104126,
      "grad_norm": 0.1265432983636856,
      "learning_rate": 2e-07,
      "loss": 0.0123,
      "step": 1785
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16669291660104182,
      "grad_norm": 0.15387628972530365,
      "learning_rate": 2e-07,
      "loss": 0.0475,
      "step": 1786
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16678624970104242,
      "grad_norm": 0.13528841733932495,
      "learning_rate": 2e-07,
      "loss": 0.0213,
      "step": 1787
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.166879582801043,
      "grad_norm": 0.1590741127729416,
      "learning_rate": 2e-07,
      "loss": 0.02,
      "step": 1788
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16697291590104357,
      "grad_norm": 0.14132629334926605,
      "learning_rate": 2e-07,
      "loss": 0.025,
      "step": 1789
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16706624900104416,
      "grad_norm": 0.13157938420772552,
      "learning_rate": 2e-07,
      "loss": -0.004,
      "step": 1790
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16715958210104476,
      "grad_norm": 0.1383260041475296,
      "learning_rate": 2e-07,
      "loss": 0.039,
      "step": 1791
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16725291520104532,
      "grad_norm": 0.12811826169490814,
      "learning_rate": 2e-07,
      "loss": 0.0213,
      "step": 1792
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1673462483010459,
      "grad_norm": 0.14032649993896484,
      "learning_rate": 2e-07,
      "loss": 0.0318,
      "step": 1793
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1674395814010465,
      "grad_norm": 0.13948988914489746,
      "learning_rate": 2e-07,
      "loss": 0.0479,
      "step": 1794
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16753291450104707,
      "grad_norm": 0.14622651040554047,
      "learning_rate": 2e-07,
      "loss": 0.0349,
      "step": 1795
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16762624760104766,
      "grad_norm": 0.14201799035072327,
      "learning_rate": 2e-07,
      "loss": 0.011,
      "step": 1796
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16771958070104825,
      "grad_norm": 0.13881975412368774,
      "learning_rate": 2e-07,
      "loss": 0.0214,
      "step": 1797
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16781291380104882,
      "grad_norm": 0.13623951375484467,
      "learning_rate": 2e-07,
      "loss": 0.0196,
      "step": 1798
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1679062469010494,
      "grad_norm": 0.15504132211208344,
      "learning_rate": 2e-07,
      "loss": 0.0214,
      "step": 1799
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16799958000105,
      "grad_norm": 0.13203254342079163,
      "learning_rate": 2e-07,
      "loss": -0.0011,
      "step": 1800
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16809291310105057,
      "grad_norm": 0.14779958128929138,
      "learning_rate": 2e-07,
      "loss": 0.0437,
      "step": 1801
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16818624620105116,
      "grad_norm": 0.13587374985218048,
      "learning_rate": 2e-07,
      "loss": 0.0465,
      "step": 1802
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16827957930105175,
      "grad_norm": 0.14497019350528717,
      "learning_rate": 2e-07,
      "loss": 0.0636,
      "step": 1803
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16837291240105234,
      "grad_norm": 0.1412143111228943,
      "learning_rate": 2e-07,
      "loss": 0.0519,
      "step": 1804
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1684662455010529,
      "grad_norm": 0.13916926085948944,
      "learning_rate": 2e-07,
      "loss": 0.0206,
      "step": 1805
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1685595786010535,
      "grad_norm": 0.13273264467716217,
      "learning_rate": 2e-07,
      "loss": 0.0105,
      "step": 1806
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1686529117010541,
      "grad_norm": 0.13299500942230225,
      "learning_rate": 2e-07,
      "loss": 0.0405,
      "step": 1807
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16874624480105466,
      "grad_norm": 0.16355222463607788,
      "learning_rate": 2e-07,
      "loss": 0.0545,
      "step": 1808
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16883957790105525,
      "grad_norm": 0.14443746209144592,
      "learning_rate": 2e-07,
      "loss": 0.0568,
      "step": 1809
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16893291100105584,
      "grad_norm": 0.13638895750045776,
      "learning_rate": 2e-07,
      "loss": 0.0095,
      "step": 1810
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1690262441010564,
      "grad_norm": 0.13642951846122742,
      "learning_rate": 2e-07,
      "loss": 0.0211,
      "step": 1811
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.169119577201057,
      "grad_norm": 0.14684197306632996,
      "learning_rate": 2e-07,
      "loss": 0.0252,
      "step": 1812
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1692129103010576,
      "grad_norm": 0.15036767721176147,
      "learning_rate": 2e-07,
      "loss": 0.0117,
      "step": 1813
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16930624340105815,
      "grad_norm": 0.12976665794849396,
      "learning_rate": 2e-07,
      "loss": 0.0508,
      "step": 1814
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16939957650105875,
      "grad_norm": 0.12477904558181763,
      "learning_rate": 2e-07,
      "loss": 0.0155,
      "step": 1815
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16949290960105934,
      "grad_norm": 0.14248622953891754,
      "learning_rate": 2e-07,
      "loss": 0.0571,
      "step": 1816
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1695862427010599,
      "grad_norm": 0.1477234810590744,
      "learning_rate": 2e-07,
      "loss": 0.007,
      "step": 1817
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1696795758010605,
      "grad_norm": 0.1388540118932724,
      "learning_rate": 2e-07,
      "loss": 0.0404,
      "step": 1818
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1697729089010611,
      "grad_norm": 0.13827331364154816,
      "learning_rate": 2e-07,
      "loss": 0.0086,
      "step": 1819
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16986624200106165,
      "grad_norm": 0.1328304260969162,
      "learning_rate": 2e-07,
      "loss": 0.0031,
      "step": 1820
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.16995957510106224,
      "grad_norm": 0.16423270106315613,
      "learning_rate": 2e-07,
      "loss": 0.0478,
      "step": 1821
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17005290820106284,
      "grad_norm": 0.17471261322498322,
      "learning_rate": 2e-07,
      "loss": 0.077,
      "step": 1822
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17014624130106343,
      "grad_norm": 0.14542202651500702,
      "learning_rate": 2e-07,
      "loss": 0.0019,
      "step": 1823
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.170239574401064,
      "grad_norm": 0.14537037909030914,
      "learning_rate": 2e-07,
      "loss": -0.0299,
      "step": 1824
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.009905133928571397,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4025.0,
      "completions/mean_length": 617.5223388671875,
      "completions/mean_terminated_length": 582.7228393554688,
      "completions/min_length": 83.0,
      "completions/min_terminated_length": 83.0,
      "epoch": 0.17033290750106458,
      "grad_norm": 0.12343145161867142,
      "learning_rate": 2e-07,
      "loss": 0.0192,
      "num_tokens": 1226943024.0,
      "reward": 0.6091831922531128,
      "reward_std": 0.18122167885303497,
      "rewards/simpleverify_reward/mean": 0.6091831922531128,
      "rewards/simpleverify_reward/std": 0.48793768882751465,
      "step": 1825
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17042624060106518,
      "grad_norm": 0.12463295459747314,
      "learning_rate": 2e-07,
      "loss": 0.1004,
      "step": 1826
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17051957370106574,
      "grad_norm": 0.12601059675216675,
      "learning_rate": 2e-07,
      "loss": 0.0171,
      "step": 1827
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17061290680106633,
      "grad_norm": 0.1419769674539566,
      "learning_rate": 2e-07,
      "loss": -0.0393,
      "step": 1828
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17070623990106693,
      "grad_norm": 0.13415008783340454,
      "learning_rate": 2e-07,
      "loss": -0.0069,
      "step": 1829
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1707995730010675,
      "grad_norm": 0.13519296050071716,
      "learning_rate": 2e-07,
      "loss": -0.0246,
      "step": 1830
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17089290610106808,
      "grad_norm": 0.13976716995239258,
      "learning_rate": 2e-07,
      "loss": 0.0529,
      "step": 1831
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17098623920106867,
      "grad_norm": 0.13812245428562164,
      "learning_rate": 2e-07,
      "loss": 0.0206,
      "step": 1832
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17107957230106924,
      "grad_norm": 0.13825631141662598,
      "learning_rate": 2e-07,
      "loss": 0.0094,
      "step": 1833
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17117290540106983,
      "grad_norm": 0.14670385420322418,
      "learning_rate": 2e-07,
      "loss": 0.041,
      "step": 1834
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17126623850107042,
      "grad_norm": 0.12002365291118622,
      "learning_rate": 2e-07,
      "loss": -0.0302,
      "step": 1835
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.171359571601071,
      "grad_norm": 0.15034233033657074,
      "learning_rate": 2e-07,
      "loss": 0.0327,
      "step": 1836
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17145290470107158,
      "grad_norm": 0.14089319109916687,
      "learning_rate": 2e-07,
      "loss": 0.0384,
      "step": 1837
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17154623780107217,
      "grad_norm": 0.1912098526954651,
      "learning_rate": 2e-07,
      "loss": 0.0373,
      "step": 1838
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17163957090107274,
      "grad_norm": 0.12161079794168472,
      "learning_rate": 2e-07,
      "loss": 0.023,
      "step": 1839
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17173290400107333,
      "grad_norm": 0.14579805731773376,
      "learning_rate": 2e-07,
      "loss": 0.0029,
      "step": 1840
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17182623710107392,
      "grad_norm": 0.12602153420448303,
      "learning_rate": 2e-07,
      "loss": 0.0902,
      "step": 1841
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17191957020107448,
      "grad_norm": 0.13122965395450592,
      "learning_rate": 2e-07,
      "loss": -0.0086,
      "step": 1842
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17201290330107508,
      "grad_norm": 0.14741212129592896,
      "learning_rate": 2e-07,
      "loss": 0.0245,
      "step": 1843
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17210623640107567,
      "grad_norm": 0.1368563175201416,
      "learning_rate": 2e-07,
      "loss": 0.0862,
      "step": 1844
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17219956950107626,
      "grad_norm": 0.13294988870620728,
      "learning_rate": 2e-07,
      "loss": 0.0676,
      "step": 1845
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17229290260107682,
      "grad_norm": 0.1595952808856964,
      "learning_rate": 2e-07,
      "loss": 0.0328,
      "step": 1846
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17238623570107742,
      "grad_norm": 0.1500539928674698,
      "learning_rate": 2e-07,
      "loss": 0.0238,
      "step": 1847
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.172479568801078,
      "grad_norm": 0.14594227075576782,
      "learning_rate": 2e-07,
      "loss": 0.0801,
      "step": 1848
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17257290190107857,
      "grad_norm": 0.13235926628112793,
      "learning_rate": 2e-07,
      "loss": 0.0286,
      "step": 1849
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17266623500107917,
      "grad_norm": 0.12428905814886093,
      "learning_rate": 2e-07,
      "loss": 0.0015,
      "step": 1850
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17275956810107976,
      "grad_norm": 0.1370844841003418,
      "learning_rate": 2e-07,
      "loss": 0.0355,
      "step": 1851
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17285290120108032,
      "grad_norm": 0.1450466364622116,
      "learning_rate": 2e-07,
      "loss": -0.0069,
      "step": 1852
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17294623430108091,
      "grad_norm": 0.15126030147075653,
      "learning_rate": 2e-07,
      "loss": -0.0009,
      "step": 1853
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1730395674010815,
      "grad_norm": 0.17182661592960358,
      "learning_rate": 2e-07,
      "loss": 0.02,
      "step": 1854
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17313290050108207,
      "grad_norm": 0.14480751752853394,
      "learning_rate": 2e-07,
      "loss": 0.0011,
      "step": 1855
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17322623360108266,
      "grad_norm": 0.16947631537914276,
      "learning_rate": 2e-07,
      "loss": 0.0086,
      "step": 1856
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17331956670108326,
      "grad_norm": 0.13966505229473114,
      "learning_rate": 2e-07,
      "loss": 0.0229,
      "step": 1857
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17341289980108382,
      "grad_norm": 0.136332705616951,
      "learning_rate": 2e-07,
      "loss": 0.0121,
      "step": 1858
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1735062329010844,
      "grad_norm": 0.1282629668712616,
      "learning_rate": 2e-07,
      "loss": 0.0134,
      "step": 1859
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.173599566001085,
      "grad_norm": 0.12012812495231628,
      "learning_rate": 2e-07,
      "loss": -0.0126,
      "step": 1860
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17369289910108557,
      "grad_norm": 0.14423710107803345,
      "learning_rate": 2e-07,
      "loss": 0.0481,
      "step": 1861
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17378623220108616,
      "grad_norm": 0.15105998516082764,
      "learning_rate": 2e-07,
      "loss": 0.0624,
      "step": 1862
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17387956530108675,
      "grad_norm": 0.1641751080751419,
      "learning_rate": 2e-07,
      "loss": 0.059,
      "step": 1863
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17397289840108732,
      "grad_norm": 0.1319325864315033,
      "learning_rate": 2e-07,
      "loss": 0.0417,
      "step": 1864
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1740662315010879,
      "grad_norm": 0.1350196897983551,
      "learning_rate": 2e-07,
      "loss": 0.0371,
      "step": 1865
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1741595646010885,
      "grad_norm": 0.14809785783290863,
      "learning_rate": 2e-07,
      "loss": 0.0168,
      "step": 1866
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1742528977010891,
      "grad_norm": 0.12649235129356384,
      "learning_rate": 2e-07,
      "loss": -0.0114,
      "step": 1867
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17434623080108966,
      "grad_norm": 0.14483027160167694,
      "learning_rate": 2e-07,
      "loss": 0.0109,
      "step": 1868
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17443956390109025,
      "grad_norm": 0.13062791526317596,
      "learning_rate": 2e-07,
      "loss": 0.009,
      "step": 1869
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17453289700109084,
      "grad_norm": 0.13055458664894104,
      "learning_rate": 2e-07,
      "loss": 0.0324,
      "step": 1870
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1746262301010914,
      "grad_norm": 0.14936086535453796,
      "learning_rate": 2e-07,
      "loss": 0.0341,
      "step": 1871
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.174719563201092,
      "grad_norm": 0.1262301653623581,
      "learning_rate": 2e-07,
      "loss": 0.0143,
      "step": 1872
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1748128963010926,
      "grad_norm": 0.13482236862182617,
      "learning_rate": 2e-07,
      "loss": -0.0176,
      "step": 1873
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17490622940109316,
      "grad_norm": 0.1613515168428421,
      "learning_rate": 2e-07,
      "loss": 0.0465,
      "step": 1874
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17499956250109375,
      "grad_norm": 0.15323977172374725,
      "learning_rate": 2e-07,
      "loss": 0.0259,
      "step": 1875
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17509289560109434,
      "grad_norm": 0.15485461056232452,
      "learning_rate": 2e-07,
      "loss": 0.0373,
      "step": 1876
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1751862287010949,
      "grad_norm": 0.13084079325199127,
      "learning_rate": 2e-07,
      "loss": 0.0189,
      "step": 1877
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1752795618010955,
      "grad_norm": 0.14444264769554138,
      "learning_rate": 2e-07,
      "loss": 0.0516,
      "step": 1878
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1753728949010961,
      "grad_norm": 0.1359085738658905,
      "learning_rate": 2e-07,
      "loss": 0.0212,
      "step": 1879
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17546622800109665,
      "grad_norm": 0.14733566343784332,
      "learning_rate": 2e-07,
      "loss": 0.0036,
      "step": 1880
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17555956110109724,
      "grad_norm": 0.1420356184244156,
      "learning_rate": 2e-07,
      "loss": 0.0256,
      "step": 1881
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17565289420109784,
      "grad_norm": 0.1636306196451187,
      "learning_rate": 2e-07,
      "loss": 0.015,
      "step": 1882
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1757462273010984,
      "grad_norm": 0.14768120646476746,
      "learning_rate": 2e-07,
      "loss": 0.0305,
      "step": 1883
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.175839560401099,
      "grad_norm": 0.16058096289634705,
      "learning_rate": 2e-07,
      "loss": -0.0201,
      "step": 1884
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17593289350109959,
      "grad_norm": 0.14790402352809906,
      "learning_rate": 2e-07,
      "loss": 0.0167,
      "step": 1885
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17602622660110018,
      "grad_norm": 0.13940365612506866,
      "learning_rate": 2e-07,
      "loss": 0.0464,
      "step": 1886
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17611955970110074,
      "grad_norm": 0.13371802866458893,
      "learning_rate": 2e-07,
      "loss": 0.0241,
      "step": 1887
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17621289280110133,
      "grad_norm": 0.14206700026988983,
      "learning_rate": 2e-07,
      "loss": -0.0048,
      "step": 1888
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.010637555803571397,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4089.0,
      "completions/mean_length": 632.0046997070312,
      "completions/mean_terminated_length": 594.760009765625,
      "completions/min_length": 108.0,
      "completions/min_terminated_length": 108.0,
      "epoch": 0.17630622590110193,
      "grad_norm": 0.12968692183494568,
      "learning_rate": 2e-07,
      "loss": 0.0453,
      "num_tokens": 1268941588.0,
      "reward": 0.6137521266937256,
      "reward_std": 0.17672185599803925,
      "rewards/simpleverify_reward/mean": 0.6137520670890808,
      "rewards/simpleverify_reward/std": 0.4868928790092468,
      "step": 1889
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1763995590011025,
      "grad_norm": 0.1270974576473236,
      "learning_rate": 2e-07,
      "loss": -0.0061,
      "step": 1890
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17649289210110308,
      "grad_norm": 0.12535491585731506,
      "learning_rate": 2e-07,
      "loss": 0.0367,
      "step": 1891
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17658622520110367,
      "grad_norm": 0.12181065231561661,
      "learning_rate": 2e-07,
      "loss": 0.0281,
      "step": 1892
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17667955830110424,
      "grad_norm": 0.13414418697357178,
      "learning_rate": 2e-07,
      "loss": 0.0629,
      "step": 1893
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17677289140110483,
      "grad_norm": 0.13094033300876617,
      "learning_rate": 2e-07,
      "loss": 0.0186,
      "step": 1894
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17686622450110542,
      "grad_norm": 0.15238836407661438,
      "learning_rate": 2e-07,
      "loss": -0.0039,
      "step": 1895
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.176959557601106,
      "grad_norm": 0.1314670443534851,
      "learning_rate": 2e-07,
      "loss": 0.005,
      "step": 1896
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17705289070110658,
      "grad_norm": 0.14410607516765594,
      "learning_rate": 2e-07,
      "loss": -0.0036,
      "step": 1897
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17714622380110717,
      "grad_norm": 0.1851828396320343,
      "learning_rate": 2e-07,
      "loss": 0.059,
      "step": 1898
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17723955690110774,
      "grad_norm": 0.12756502628326416,
      "learning_rate": 2e-07,
      "loss": -0.01,
      "step": 1899
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17733289000110833,
      "grad_norm": 0.12007372826337814,
      "learning_rate": 2e-07,
      "loss": 0.0301,
      "step": 1900
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17742622310110892,
      "grad_norm": 0.1303112953901291,
      "learning_rate": 2e-07,
      "loss": 0.0174,
      "step": 1901
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17751955620110949,
      "grad_norm": 0.1678982675075531,
      "learning_rate": 2e-07,
      "loss": 0.0266,
      "step": 1902
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17761288930111008,
      "grad_norm": 0.1291574388742447,
      "learning_rate": 2e-07,
      "loss": 0.0172,
      "step": 1903
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17770622240111067,
      "grad_norm": 0.1233028993010521,
      "learning_rate": 2e-07,
      "loss": -0.0103,
      "step": 1904
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17779955550111123,
      "grad_norm": 0.13970455527305603,
      "learning_rate": 2e-07,
      "loss": -0.0202,
      "step": 1905
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17789288860111183,
      "grad_norm": 0.13417783379554749,
      "learning_rate": 2e-07,
      "loss": 0.0313,
      "step": 1906
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17798622170111242,
      "grad_norm": 0.1270092874765396,
      "learning_rate": 2e-07,
      "loss": 0.0073,
      "step": 1907
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.178079554801113,
      "grad_norm": 0.14816221594810486,
      "learning_rate": 2e-07,
      "loss": 0.0441,
      "step": 1908
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17817288790111357,
      "grad_norm": 0.139995738863945,
      "learning_rate": 2e-07,
      "loss": 0.0535,
      "step": 1909
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17826622100111417,
      "grad_norm": 0.13110674917697906,
      "learning_rate": 2e-07,
      "loss": 0.014,
      "step": 1910
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17835955410111476,
      "grad_norm": 0.14220739901065826,
      "learning_rate": 2e-07,
      "loss": 0.0105,
      "step": 1911
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17845288720111532,
      "grad_norm": 0.1262693703174591,
      "learning_rate": 2e-07,
      "loss": 0.0082,
      "step": 1912
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17854622030111592,
      "grad_norm": 0.13542824983596802,
      "learning_rate": 2e-07,
      "loss": 0.0486,
      "step": 1913
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1786395534011165,
      "grad_norm": 0.13898195326328278,
      "learning_rate": 2e-07,
      "loss": 0.0035,
      "step": 1914
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17873288650111707,
      "grad_norm": 0.1645851880311966,
      "learning_rate": 2e-07,
      "loss": 0.0194,
      "step": 1915
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17882621960111766,
      "grad_norm": 0.13773660361766815,
      "learning_rate": 2e-07,
      "loss": 0.0151,
      "step": 1916
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17891955270111826,
      "grad_norm": 0.1356283575296402,
      "learning_rate": 2e-07,
      "loss": -0.0031,
      "step": 1917
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17901288580111882,
      "grad_norm": 0.11995309591293335,
      "learning_rate": 2e-07,
      "loss": 0.0519,
      "step": 1918
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1791062189011194,
      "grad_norm": 0.13299036026000977,
      "learning_rate": 2e-07,
      "loss": 0.0273,
      "step": 1919
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17919955200112,
      "grad_norm": 0.13857421278953552,
      "learning_rate": 2e-07,
      "loss": 0.0211,
      "step": 1920
    },
    {
      "clip_ratio/high_max": 0.01674611441558227,
      "clip_ratio/high_mean": 0.007233279859065078,
      "clip_ratio/low_mean": 0.0036107119158259593,
      "clip_ratio/low_min": 0.0003255704759794753,
      "clip_ratio/region_mean": 0.010843991651199758,
      "completions/clipped_ratio": 0.008335658482142905,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4085.0,
      "completions/mean_length": 619.3406372070312,
      "completions/mean_terminated_length": 590.1167602539062,
      "completions/min_length": 90.0,
      "completions/min_terminated_length": 90.0,
      "epoch": 0.17929288510112057,
      "grad_norm": 0.09886615723371506,
      "learning_rate": 2e-07,
      "loss": 0.0146,
      "num_tokens": 1310170984.0,
      "reward": 0.6195417642593384,
      "reward_std": 0.17761610448360443,
      "rewards/simpleverify_reward/mean": 0.6195417046546936,
      "rewards/simpleverify_reward/std": 0.48550376296043396,
      "step": 1921
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17938621820112116,
      "grad_norm": 0.12264218926429749,
      "learning_rate": 2e-07,
      "loss": 0.0389,
      "step": 1922
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17947955130112175,
      "grad_norm": 0.13081879913806915,
      "learning_rate": 2e-07,
      "loss": 0.0346,
      "step": 1923
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17957288440112232,
      "grad_norm": 0.1340312957763672,
      "learning_rate": 2e-07,
      "loss": 0.0165,
      "step": 1924
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1796662175011229,
      "grad_norm": 0.1363367736339569,
      "learning_rate": 2e-07,
      "loss": 0.0437,
      "step": 1925
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1797595506011235,
      "grad_norm": 0.14061513543128967,
      "learning_rate": 2e-07,
      "loss": 0.0012,
      "step": 1926
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1798528837011241,
      "grad_norm": 0.13663895428180695,
      "learning_rate": 2e-07,
      "loss": -0.0332,
      "step": 1927
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.17994621680112466,
      "grad_norm": 0.13007336854934692,
      "learning_rate": 2e-07,
      "loss": 0.0286,
      "step": 1928
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18003954990112525,
      "grad_norm": 0.12456807494163513,
      "learning_rate": 2e-07,
      "loss": -0.0047,
      "step": 1929
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18013288300112584,
      "grad_norm": 0.14594756066799164,
      "learning_rate": 2e-07,
      "loss": 0.028,
      "step": 1930
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1802262161011264,
      "grad_norm": 0.1262948215007782,
      "learning_rate": 2e-07,
      "loss": -0.0199,
      "step": 1931
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.180319549201127,
      "grad_norm": 0.12121875584125519,
      "learning_rate": 2e-07,
      "loss": 0.01,
      "step": 1932
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1804128823011276,
      "grad_norm": 0.1400436908006668,
      "learning_rate": 2e-07,
      "loss": 0.0231,
      "step": 1933
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18050621540112816,
      "grad_norm": 0.13571596145629883,
      "learning_rate": 2e-07,
      "loss": -0.0381,
      "step": 1934
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18059954850112875,
      "grad_norm": 0.12717264890670776,
      "learning_rate": 2e-07,
      "loss": 0.0003,
      "step": 1935
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18069288160112934,
      "grad_norm": 0.11864596605300903,
      "learning_rate": 2e-07,
      "loss": 0.0501,
      "step": 1936
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1807862147011299,
      "grad_norm": 0.16038836538791656,
      "learning_rate": 2e-07,
      "loss": 0.051,
      "step": 1937
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1808795478011305,
      "grad_norm": 0.12937262654304504,
      "learning_rate": 2e-07,
      "loss": 0.0376,
      "step": 1938
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1809728809011311,
      "grad_norm": 0.1536262035369873,
      "learning_rate": 2e-07,
      "loss": 0.0422,
      "step": 1939
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18106621400113165,
      "grad_norm": 0.12450678646564484,
      "learning_rate": 2e-07,
      "loss": 0.0078,
      "step": 1940
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18115954710113225,
      "grad_norm": 0.1392851024866104,
      "learning_rate": 2e-07,
      "loss": 0.0366,
      "step": 1941
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18125288020113284,
      "grad_norm": 0.12198419123888016,
      "learning_rate": 2e-07,
      "loss": -0.0018,
      "step": 1942
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1813462133011334,
      "grad_norm": 0.1918262392282486,
      "learning_rate": 2e-07,
      "loss": 0.049,
      "step": 1943
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.181439546401134,
      "grad_norm": 0.13931629061698914,
      "learning_rate": 2e-07,
      "loss": 0.0782,
      "step": 1944
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1815328795011346,
      "grad_norm": 0.15207862854003906,
      "learning_rate": 2e-07,
      "loss": 0.0325,
      "step": 1945
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18162621260113515,
      "grad_norm": 0.11981978267431259,
      "learning_rate": 2e-07,
      "loss": 0.0334,
      "step": 1946
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18171954570113574,
      "grad_norm": 0.14882461726665497,
      "learning_rate": 2e-07,
      "loss": 0.0173,
      "step": 1947
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18181287880113634,
      "grad_norm": 0.13769622147083282,
      "learning_rate": 2e-07,
      "loss": 0.0025,
      "step": 1948
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18190621190113693,
      "grad_norm": 0.18945784866809845,
      "learning_rate": 2e-07,
      "loss": 0.0167,
      "step": 1949
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1819995450011375,
      "grad_norm": 0.1421399563550949,
      "learning_rate": 2e-07,
      "loss": -0.0259,
      "step": 1950
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18209287810113808,
      "grad_norm": 0.13331574201583862,
      "learning_rate": 2e-07,
      "loss": -0.0295,
      "step": 1951
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18218621120113868,
      "grad_norm": 0.14329425990581512,
      "learning_rate": 2e-07,
      "loss": 0.0432,
      "step": 1952
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18227954430113924,
      "grad_norm": 0.14803071320056915,
      "learning_rate": 2e-07,
      "loss": 0.013,
      "step": 1953
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18237287740113983,
      "grad_norm": 0.16635844111442566,
      "learning_rate": 2e-07,
      "loss": 0.0408,
      "step": 1954
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18246621050114042,
      "grad_norm": 0.13200974464416504,
      "learning_rate": 2e-07,
      "loss": 0.0292,
      "step": 1955
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.182559543601141,
      "grad_norm": 0.1481582671403885,
      "learning_rate": 2e-07,
      "loss": 0.0122,
      "step": 1956
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18265287670114158,
      "grad_norm": 0.1292586624622345,
      "learning_rate": 2e-07,
      "loss": 0.0407,
      "step": 1957
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18274620980114217,
      "grad_norm": 0.13599826395511627,
      "learning_rate": 2e-07,
      "loss": 0.0012,
      "step": 1958
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18283954290114274,
      "grad_norm": 0.14243608713150024,
      "learning_rate": 2e-07,
      "loss": 0.0235,
      "step": 1959
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18293287600114333,
      "grad_norm": 0.12822553515434265,
      "learning_rate": 2e-07,
      "loss": 0.0219,
      "step": 1960
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18302620910114392,
      "grad_norm": 0.1303199827671051,
      "learning_rate": 2e-07,
      "loss": 0.0244,
      "step": 1961
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1831195422011445,
      "grad_norm": 0.1413554698228836,
      "learning_rate": 2e-07,
      "loss": 0.0261,
      "step": 1962
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18321287530114508,
      "grad_norm": 0.14116507768630981,
      "learning_rate": 2e-07,
      "loss": -0.0197,
      "step": 1963
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18330620840114567,
      "grad_norm": 0.11513558775186539,
      "learning_rate": 2e-07,
      "loss": 0.0384,
      "step": 1964
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18339954150114623,
      "grad_norm": 0.1423267126083374,
      "learning_rate": 2e-07,
      "loss": 0.0146,
      "step": 1965
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18349287460114683,
      "grad_norm": 0.1290888786315918,
      "learning_rate": 2e-07,
      "loss": -0.0045,
      "step": 1966
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18358620770114742,
      "grad_norm": 0.1253574788570404,
      "learning_rate": 2e-07,
      "loss": 0.0003,
      "step": 1967
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18367954080114798,
      "grad_norm": 0.15042288601398468,
      "learning_rate": 2e-07,
      "loss": 0.0353,
      "step": 1968
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18377287390114858,
      "grad_norm": 0.1410323530435562,
      "learning_rate": 2e-07,
      "loss": 0.0099,
      "step": 1969
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18386620700114917,
      "grad_norm": 0.1407528519630432,
      "learning_rate": 2e-07,
      "loss": 0.0243,
      "step": 1970
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18395954010114976,
      "grad_norm": 0.14316454529762268,
      "learning_rate": 2e-07,
      "loss": 0.074,
      "step": 1971
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18405287320115032,
      "grad_norm": 0.13648678362369537,
      "learning_rate": 2e-07,
      "loss": 0.0411,
      "step": 1972
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18414620630115092,
      "grad_norm": 0.16273942589759827,
      "learning_rate": 2e-07,
      "loss": 0.0054,
      "step": 1973
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1842395394011515,
      "grad_norm": 0.14621920883655548,
      "learning_rate": 2e-07,
      "loss": 0.0527,
      "step": 1974
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18433287250115207,
      "grad_norm": 0.17995776236057281,
      "learning_rate": 2e-07,
      "loss": -0.01,
      "step": 1975
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18442620560115267,
      "grad_norm": 0.1395529806613922,
      "learning_rate": 2e-07,
      "loss": 0.0605,
      "step": 1976
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18451953870115326,
      "grad_norm": 0.136857271194458,
      "learning_rate": 2e-07,
      "loss": 0.0215,
      "step": 1977
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18461287180115382,
      "grad_norm": 0.13114769756793976,
      "learning_rate": 2e-07,
      "loss": 0.0209,
      "step": 1978
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1847062049011544,
      "grad_norm": 0.1532019078731537,
      "learning_rate": 2e-07,
      "loss": 0.057,
      "step": 1979
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.184799538001155,
      "grad_norm": 0.13745979964733124,
      "learning_rate": 2e-07,
      "loss": 0.0009,
      "step": 1980
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18489287110115557,
      "grad_norm": 0.14160163700580597,
      "learning_rate": 2e-07,
      "loss": 0.0729,
      "step": 1981
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18498620420115616,
      "grad_norm": 0.1361771821975708,
      "learning_rate": 2e-07,
      "loss": 0.0142,
      "step": 1982
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18507953730115675,
      "grad_norm": 0.14404182136058807,
      "learning_rate": 2e-07,
      "loss": 0.0106,
      "step": 1983
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18517287040115732,
      "grad_norm": 0.13711047172546387,
      "learning_rate": 2e-07,
      "loss": -0.0038,
      "step": 1984
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.009381975446428603,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4061.0,
      "completions/mean_length": 626.4524536132812,
      "completions/mean_terminated_length": 593.5928344726562,
      "completions/min_length": 92.0,
      "completions/min_terminated_length": 92.0,
      "epoch": 0.1852662035011579,
      "grad_norm": 0.12949836254119873,
      "learning_rate": 2e-07,
      "loss": -0.0001,
      "num_tokens": 1351795941.0,
      "reward": 0.6247907876968384,
      "reward_std": 0.17726653814315796,
      "rewards/simpleverify_reward/mean": 0.6247907280921936,
      "rewards/simpleverify_reward/std": 0.48418107628822327,
      "step": 1985
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1853595366011585,
      "grad_norm": 0.13212190568447113,
      "learning_rate": 2e-07,
      "loss": -0.0413,
      "step": 1986
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18545286970115907,
      "grad_norm": 0.14804911613464355,
      "learning_rate": 2e-07,
      "loss": 0.0191,
      "step": 1987
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18554620280115966,
      "grad_norm": 0.13002555072307587,
      "learning_rate": 2e-07,
      "loss": -0.0198,
      "step": 1988
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18563953590116025,
      "grad_norm": 0.14281854033470154,
      "learning_rate": 2e-07,
      "loss": 0.0198,
      "step": 1989
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18573286900116084,
      "grad_norm": 0.12832076847553253,
      "learning_rate": 2e-07,
      "loss": 0.0755,
      "step": 1990
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1858262021011614,
      "grad_norm": 0.12089355289936066,
      "learning_rate": 2e-07,
      "loss": 0.0357,
      "step": 1991
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.185919535201162,
      "grad_norm": 0.14158028364181519,
      "learning_rate": 2e-07,
      "loss": -0.0045,
      "step": 1992
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1860128683011626,
      "grad_norm": 0.12692396342754364,
      "learning_rate": 2e-07,
      "loss": 0.0369,
      "step": 1993
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18610620140116316,
      "grad_norm": 0.1321793794631958,
      "learning_rate": 2e-07,
      "loss": 0.0077,
      "step": 1994
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18619953450116375,
      "grad_norm": 0.12042239308357239,
      "learning_rate": 2e-07,
      "loss": -0.0038,
      "step": 1995
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18629286760116434,
      "grad_norm": 0.1490997076034546,
      "learning_rate": 2e-07,
      "loss": 0.0656,
      "step": 1996
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1863862007011649,
      "grad_norm": 0.13332825899124146,
      "learning_rate": 2e-07,
      "loss": 0.0277,
      "step": 1997
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1864795338011655,
      "grad_norm": 0.13467949628829956,
      "learning_rate": 2e-07,
      "loss": 0.0452,
      "step": 1998
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1865728669011661,
      "grad_norm": 0.13089565932750702,
      "learning_rate": 2e-07,
      "loss": 0.0751,
      "step": 1999
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18666620000116665,
      "grad_norm": 0.14168663322925568,
      "learning_rate": 2e-07,
      "loss": 0.0361,
      "step": 2000
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18675953310116725,
      "grad_norm": 0.13525070250034332,
      "learning_rate": 2e-07,
      "loss": 0.066,
      "step": 2001
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18685286620116784,
      "grad_norm": 0.13899831473827362,
      "learning_rate": 2e-07,
      "loss": 0.0526,
      "step": 2002
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1869461993011684,
      "grad_norm": 0.1302626132965088,
      "learning_rate": 2e-07,
      "loss": 0.0542,
      "step": 2003
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.187039532401169,
      "grad_norm": 0.1303328573703766,
      "learning_rate": 2e-07,
      "loss": 0.0404,
      "step": 2004
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1871328655011696,
      "grad_norm": 0.11603037267923355,
      "learning_rate": 2e-07,
      "loss": 0.0121,
      "step": 2005
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18722619860117015,
      "grad_norm": 0.16481482982635498,
      "learning_rate": 2e-07,
      "loss": -0.0204,
      "step": 2006
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18731953170117074,
      "grad_norm": 0.13323041796684265,
      "learning_rate": 2e-07,
      "loss": 0.0204,
      "step": 2007
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18741286480117134,
      "grad_norm": 0.13928939402103424,
      "learning_rate": 2e-07,
      "loss": 0.0451,
      "step": 2008
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1875061979011719,
      "grad_norm": 0.1318858414888382,
      "learning_rate": 2e-07,
      "loss": 0.0075,
      "step": 2009
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1875995310011725,
      "grad_norm": 0.12346776574850082,
      "learning_rate": 2e-07,
      "loss": -0.0176,
      "step": 2010
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18769286410117308,
      "grad_norm": 0.1380702406167984,
      "learning_rate": 2e-07,
      "loss": -0.0018,
      "step": 2011
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18778619720117368,
      "grad_norm": 0.1435062736272812,
      "learning_rate": 2e-07,
      "loss": 0.0031,
      "step": 2012
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18787953030117424,
      "grad_norm": 0.11607632040977478,
      "learning_rate": 2e-07,
      "loss": 0.034,
      "step": 2013
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18797286340117483,
      "grad_norm": 0.14182893931865692,
      "learning_rate": 2e-07,
      "loss": -0.0043,
      "step": 2014
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18806619650117543,
      "grad_norm": 0.127472922205925,
      "learning_rate": 2e-07,
      "loss": 0.0377,
      "step": 2015
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.188159529601176,
      "grad_norm": 0.13234007358551025,
      "learning_rate": 2e-07,
      "loss": 0.0441,
      "step": 2016
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18825286270117658,
      "grad_norm": 0.1482890546321869,
      "learning_rate": 2e-07,
      "loss": 0.0209,
      "step": 2017
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18834619580117717,
      "grad_norm": 0.13471786677837372,
      "learning_rate": 2e-07,
      "loss": 0.0245,
      "step": 2018
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18843952890117774,
      "grad_norm": 0.1459258496761322,
      "learning_rate": 2e-07,
      "loss": -0.0077,
      "step": 2019
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18853286200117833,
      "grad_norm": 0.1374923288822174,
      "learning_rate": 2e-07,
      "loss": 0.0266,
      "step": 2020
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18862619510117892,
      "grad_norm": 0.15669749677181244,
      "learning_rate": 2e-07,
      "loss": 0.0346,
      "step": 2021
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1887195282011795,
      "grad_norm": 0.13293330371379852,
      "learning_rate": 2e-07,
      "loss": 0.009,
      "step": 2022
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18881286130118008,
      "grad_norm": 0.13252577185630798,
      "learning_rate": 2e-07,
      "loss": 0.0119,
      "step": 2023
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18890619440118067,
      "grad_norm": 0.15301652252674103,
      "learning_rate": 2e-07,
      "loss": 0.0378,
      "step": 2024
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18899952750118124,
      "grad_norm": 0.22243423759937286,
      "learning_rate": 2e-07,
      "loss": 0.0204,
      "step": 2025
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18909286060118183,
      "grad_norm": 0.2164534330368042,
      "learning_rate": 2e-07,
      "loss": 0.0202,
      "step": 2026
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18918619370118242,
      "grad_norm": 0.11972451210021973,
      "learning_rate": 2e-07,
      "loss": 0.059,
      "step": 2027
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18927952680118298,
      "grad_norm": 0.20433871448040009,
      "learning_rate": 2e-07,
      "loss": -0.0025,
      "step": 2028
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18937285990118358,
      "grad_norm": 0.13868603110313416,
      "learning_rate": 2e-07,
      "loss": 0.055,
      "step": 2029
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18946619300118417,
      "grad_norm": 0.12149716168642044,
      "learning_rate": 2e-07,
      "loss": 0.0281,
      "step": 2030
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18955952610118473,
      "grad_norm": 0.13556385040283203,
      "learning_rate": 2e-07,
      "loss": -0.0038,
      "step": 2031
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18965285920118533,
      "grad_norm": 0.12901639938354492,
      "learning_rate": 2e-07,
      "loss": 0.0028,
      "step": 2032
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18974619230118592,
      "grad_norm": 0.14690649509429932,
      "learning_rate": 2e-07,
      "loss": -0.0078,
      "step": 2033
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1898395254011865,
      "grad_norm": 0.12696930766105652,
      "learning_rate": 2e-07,
      "loss": 0.0937,
      "step": 2034
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.18993285850118707,
      "grad_norm": 0.13723741471767426,
      "learning_rate": 2e-07,
      "loss": 0.0676,
      "step": 2035
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19002619160118767,
      "grad_norm": 0.2401839941740036,
      "learning_rate": 2e-07,
      "loss": 0.0778,
      "step": 2036
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19011952470118826,
      "grad_norm": 0.13668695092201233,
      "learning_rate": 2e-07,
      "loss": 0.0071,
      "step": 2037
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19021285780118882,
      "grad_norm": 0.14171245694160461,
      "learning_rate": 2e-07,
      "loss": -0.0104,
      "step": 2038
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19030619090118941,
      "grad_norm": 0.13272352516651154,
      "learning_rate": 2e-07,
      "loss": -0.0116,
      "step": 2039
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19039952400119,
      "grad_norm": 0.17398177087306976,
      "learning_rate": 2e-07,
      "loss": 0.0223,
      "step": 2040
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19049285710119057,
      "grad_norm": 0.11527741700410843,
      "learning_rate": 2e-07,
      "loss": -0.0561,
      "step": 2041
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19058619020119116,
      "grad_norm": 0.1422957479953766,
      "learning_rate": 2e-07,
      "loss": 0.0371,
      "step": 2042
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19067952330119176,
      "grad_norm": 0.12242025136947632,
      "learning_rate": 2e-07,
      "loss": 0.0418,
      "step": 2043
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19077285640119232,
      "grad_norm": 0.1306590735912323,
      "learning_rate": 2e-07,
      "loss": 0.0283,
      "step": 2044
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1908661895011929,
      "grad_norm": 0.12630999088287354,
      "learning_rate": 2e-07,
      "loss": 0.0125,
      "step": 2045
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1909595226011935,
      "grad_norm": 0.14573551714420319,
      "learning_rate": 2e-07,
      "loss": 0.0296,
      "step": 2046
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19105285570119407,
      "grad_norm": 0.12549155950546265,
      "learning_rate": 2e-07,
      "loss": 0.0583,
      "step": 2047
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19114618880119466,
      "grad_norm": 0.12897123396396637,
      "learning_rate": 2e-07,
      "loss": 0.0142,
      "step": 2048
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.008980887276785698,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4084.0,
      "completions/mean_length": 621.1405029296875,
      "completions/mean_terminated_length": 589.6502685546875,
      "completions/min_length": 74.0,
      "completions/min_terminated_length": 74.0,
      "epoch": 0.19123952190119525,
      "grad_norm": 0.13816216588020325,
      "learning_rate": 2e-07,
      "loss": 0.0352,
      "num_tokens": 1393137939.0,
      "reward": 0.6235700845718384,
      "reward_std": 0.17483706772327423,
      "rewards/simpleverify_reward/mean": 0.6235700249671936,
      "rewards/simpleverify_reward/std": 0.48449406027793884,
      "step": 2049
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19133285500119582,
      "grad_norm": 0.13441555202007294,
      "learning_rate": 2e-07,
      "loss": 0.0308,
      "step": 2050
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1914261881011964,
      "grad_norm": 0.1378239244222641,
      "learning_rate": 2e-07,
      "loss": -0.0108,
      "step": 2051
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.191519521201197,
      "grad_norm": 0.14629487693309784,
      "learning_rate": 2e-07,
      "loss": 0.035,
      "step": 2052
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1916128543011976,
      "grad_norm": 0.12785065174102783,
      "learning_rate": 2e-07,
      "loss": 0.0105,
      "step": 2053
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19170618740119816,
      "grad_norm": 0.13275893032550812,
      "learning_rate": 2e-07,
      "loss": 0.0131,
      "step": 2054
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19179952050119875,
      "grad_norm": 0.13304942846298218,
      "learning_rate": 2e-07,
      "loss": -0.0095,
      "step": 2055
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19189285360119934,
      "grad_norm": 0.12484243512153625,
      "learning_rate": 2e-07,
      "loss": 0.0526,
      "step": 2056
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1919861867011999,
      "grad_norm": 0.13004404306411743,
      "learning_rate": 2e-07,
      "loss": 0.0123,
      "step": 2057
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1920795198012005,
      "grad_norm": 0.2022404670715332,
      "learning_rate": 2e-07,
      "loss": -0.003,
      "step": 2058
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1921728529012011,
      "grad_norm": 0.14473304152488708,
      "learning_rate": 2e-07,
      "loss": 0.0643,
      "step": 2059
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19226618600120166,
      "grad_norm": 0.14179791510105133,
      "learning_rate": 2e-07,
      "loss": -0.0284,
      "step": 2060
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19235951910120225,
      "grad_norm": 0.1339181810617447,
      "learning_rate": 2e-07,
      "loss": -0.0113,
      "step": 2061
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19245285220120284,
      "grad_norm": 0.1328258365392685,
      "learning_rate": 2e-07,
      "loss": -0.0076,
      "step": 2062
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1925461853012034,
      "grad_norm": 0.14393477141857147,
      "learning_rate": 2e-07,
      "loss": 0.0397,
      "step": 2063
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.192639518401204,
      "grad_norm": 0.12550607323646545,
      "learning_rate": 2e-07,
      "loss": 0.024,
      "step": 2064
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1927328515012046,
      "grad_norm": 0.13866785168647766,
      "learning_rate": 2e-07,
      "loss": 0.0372,
      "step": 2065
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19282618460120515,
      "grad_norm": 0.13745541870594025,
      "learning_rate": 2e-07,
      "loss": 0.0199,
      "step": 2066
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19291951770120574,
      "grad_norm": 0.21414770185947418,
      "learning_rate": 2e-07,
      "loss": 0.0039,
      "step": 2067
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19301285080120634,
      "grad_norm": 0.12601198256015778,
      "learning_rate": 2e-07,
      "loss": 0.0283,
      "step": 2068
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1931061839012069,
      "grad_norm": 0.13887163996696472,
      "learning_rate": 2e-07,
      "loss": 0.0486,
      "step": 2069
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1931995170012075,
      "grad_norm": 0.12245846539735794,
      "learning_rate": 2e-07,
      "loss": 0.0181,
      "step": 2070
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19329285010120809,
      "grad_norm": 0.1408935785293579,
      "learning_rate": 2e-07,
      "loss": 0.0514,
      "step": 2071
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19338618320120865,
      "grad_norm": 0.13606879115104675,
      "learning_rate": 2e-07,
      "loss": 0.0526,
      "step": 2072
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19347951630120924,
      "grad_norm": 0.1314115673303604,
      "learning_rate": 2e-07,
      "loss": 0.0333,
      "step": 2073
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19357284940120983,
      "grad_norm": 0.1365850865840912,
      "learning_rate": 2e-07,
      "loss": 0.0171,
      "step": 2074
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19366618250121043,
      "grad_norm": 0.11020214110612869,
      "learning_rate": 2e-07,
      "loss": 0.0461,
      "step": 2075
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.193759515601211,
      "grad_norm": 0.42245760560035706,
      "learning_rate": 2e-07,
      "loss": 0.0189,
      "step": 2076
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19385284870121158,
      "grad_norm": 0.17467431724071503,
      "learning_rate": 2e-07,
      "loss": 0.011,
      "step": 2077
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19394618180121218,
      "grad_norm": 0.12905199825763702,
      "learning_rate": 2e-07,
      "loss": 0.0286,
      "step": 2078
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19403951490121274,
      "grad_norm": 0.1307622492313385,
      "learning_rate": 2e-07,
      "loss": -0.0261,
      "step": 2079
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19413284800121333,
      "grad_norm": 0.12891347706317902,
      "learning_rate": 2e-07,
      "loss": 0.0312,
      "step": 2080
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19422618110121392,
      "grad_norm": 0.16293098032474518,
      "learning_rate": 2e-07,
      "loss": 0.0479,
      "step": 2081
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1943195142012145,
      "grad_norm": 0.13904277980327606,
      "learning_rate": 2e-07,
      "loss": 0.0098,
      "step": 2082
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19441284730121508,
      "grad_norm": 0.14107856154441833,
      "learning_rate": 2e-07,
      "loss": 0.0439,
      "step": 2083
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19450618040121567,
      "grad_norm": 0.13813064992427826,
      "learning_rate": 2e-07,
      "loss": 0.0029,
      "step": 2084
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19459951350121624,
      "grad_norm": 0.1166517436504364,
      "learning_rate": 2e-07,
      "loss": 0.0467,
      "step": 2085
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19469284660121683,
      "grad_norm": 0.1782504916191101,
      "learning_rate": 2e-07,
      "loss": -0.0165,
      "step": 2086
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19478617970121742,
      "grad_norm": 0.13020426034927368,
      "learning_rate": 2e-07,
      "loss": 0.0333,
      "step": 2087
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19487951280121799,
      "grad_norm": 0.1313367635011673,
      "learning_rate": 2e-07,
      "loss": 0.0231,
      "step": 2088
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19497284590121858,
      "grad_norm": 0.1328446865081787,
      "learning_rate": 2e-07,
      "loss": 0.0104,
      "step": 2089
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19506617900121917,
      "grad_norm": 0.14593270421028137,
      "learning_rate": 2e-07,
      "loss": 0.0168,
      "step": 2090
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19515951210121973,
      "grad_norm": 0.1278373748064041,
      "learning_rate": 2e-07,
      "loss": 0.0039,
      "step": 2091
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19525284520122033,
      "grad_norm": 0.13942936062812805,
      "learning_rate": 2e-07,
      "loss": 0.0317,
      "step": 2092
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19534617830122092,
      "grad_norm": 0.1356835663318634,
      "learning_rate": 2e-07,
      "loss": 0.0623,
      "step": 2093
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1954395114012215,
      "grad_norm": 0.14816170930862427,
      "learning_rate": 2e-07,
      "loss": 0.0392,
      "step": 2094
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19553284450122208,
      "grad_norm": 0.1527928113937378,
      "learning_rate": 2e-07,
      "loss": 0.0111,
      "step": 2095
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19562617760122267,
      "grad_norm": 0.11939983814954758,
      "learning_rate": 2e-07,
      "loss": -0.0336,
      "step": 2096
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19571951070122326,
      "grad_norm": 0.12709404528141022,
      "learning_rate": 2e-07,
      "loss": 0.0484,
      "step": 2097
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19581284380122382,
      "grad_norm": 0.1341167837381363,
      "learning_rate": 2e-07,
      "loss": 0.0576,
      "step": 2098
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19590617690122442,
      "grad_norm": 0.11741939932107925,
      "learning_rate": 2e-07,
      "loss": -0.0014,
      "step": 2099
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.195999510001225,
      "grad_norm": 0.155959814786911,
      "learning_rate": 2e-07,
      "loss": 0.0223,
      "step": 2100
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19609284310122557,
      "grad_norm": 0.1319066435098648,
      "learning_rate": 2e-07,
      "loss": 0.0044,
      "step": 2101
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19618617620122616,
      "grad_norm": 0.12806671857833862,
      "learning_rate": 2e-07,
      "loss": 0.036,
      "step": 2102
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19627950930122676,
      "grad_norm": 0.13918417692184448,
      "learning_rate": 2e-07,
      "loss": 0.0182,
      "step": 2103
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19637284240122732,
      "grad_norm": 0.13889160752296448,
      "learning_rate": 2e-07,
      "loss": -0.0009,
      "step": 2104
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1964661755012279,
      "grad_norm": 0.1448560208082199,
      "learning_rate": 2e-07,
      "loss": 0.0411,
      "step": 2105
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1965595086012285,
      "grad_norm": 0.12901759147644043,
      "learning_rate": 2e-07,
      "loss": 0.0262,
      "step": 2106
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19665284170122907,
      "grad_norm": 0.14847217500209808,
      "learning_rate": 2e-07,
      "loss": 0.0547,
      "step": 2107
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19674617480122966,
      "grad_norm": 0.17445239424705505,
      "learning_rate": 2e-07,
      "loss": 0.0248,
      "step": 2108
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19683950790123025,
      "grad_norm": 0.12774260342121124,
      "learning_rate": 2e-07,
      "loss": 0.0214,
      "step": 2109
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19693284100123082,
      "grad_norm": 0.12575243413448334,
      "learning_rate": 2e-07,
      "loss": -0.0035,
      "step": 2110
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1970261741012314,
      "grad_norm": 0.14290933310985565,
      "learning_rate": 2e-07,
      "loss": 0.0552,
      "step": 2111
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.197119507201232,
      "grad_norm": 0.14731821417808533,
      "learning_rate": 2e-07,
      "loss": 0.0309,
      "step": 2112
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.009852818080357095,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4087.0,
      "completions/mean_length": 629.9486083984375,
      "completions/mean_terminated_length": 595.4583740234375,
      "completions/min_length": 83.0,
      "completions/min_terminated_length": 83.0,
      "epoch": 0.19721284030123257,
      "grad_norm": 0.13937370479106903,
      "learning_rate": 2e-07,
      "loss": 0.0692,
      "num_tokens": 1435032328.0,
      "reward": 0.6246512532234192,
      "reward_std": 0.1760890781879425,
      "rewards/simpleverify_reward/mean": 0.6246512532234192,
      "rewards/simpleverify_reward/std": 0.48421710729599,
      "step": 2113
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19730617340123316,
      "grad_norm": 0.13251301646232605,
      "learning_rate": 2e-07,
      "loss": 0.041,
      "step": 2114
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19739950650123375,
      "grad_norm": 0.13234274089336395,
      "learning_rate": 2e-07,
      "loss": -0.0099,
      "step": 2115
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19749283960123434,
      "grad_norm": 0.1307438164949417,
      "learning_rate": 2e-07,
      "loss": 0.0473,
      "step": 2116
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1975861727012349,
      "grad_norm": 0.13851957023143768,
      "learning_rate": 2e-07,
      "loss": -0.0179,
      "step": 2117
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1976795058012355,
      "grad_norm": 0.10829498618841171,
      "learning_rate": 2e-07,
      "loss": 0.0232,
      "step": 2118
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1977728389012361,
      "grad_norm": 0.14571848511695862,
      "learning_rate": 2e-07,
      "loss": 0.0283,
      "step": 2119
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19786617200123666,
      "grad_norm": 0.12630069255828857,
      "learning_rate": 2e-07,
      "loss": 0.0295,
      "step": 2120
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19795950510123725,
      "grad_norm": 0.11628395318984985,
      "learning_rate": 2e-07,
      "loss": 0.0504,
      "step": 2121
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19805283820123784,
      "grad_norm": 0.13721053302288055,
      "learning_rate": 2e-07,
      "loss": 0.0328,
      "step": 2122
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1981461713012384,
      "grad_norm": 0.12931771576404572,
      "learning_rate": 2e-07,
      "loss": 0.0114,
      "step": 2123
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.198239504401239,
      "grad_norm": 0.12818796932697296,
      "learning_rate": 2e-07,
      "loss": 0.0193,
      "step": 2124
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1983328375012396,
      "grad_norm": 0.12518827617168427,
      "learning_rate": 2e-07,
      "loss": 0.0132,
      "step": 2125
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19842617060124015,
      "grad_norm": 0.1292812079191208,
      "learning_rate": 2e-07,
      "loss": 0.016,
      "step": 2126
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19851950370124075,
      "grad_norm": 0.12755903601646423,
      "learning_rate": 2e-07,
      "loss": 0.0367,
      "step": 2127
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19861283680124134,
      "grad_norm": 0.14565172791481018,
      "learning_rate": 2e-07,
      "loss": 0.0083,
      "step": 2128
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1987061699012419,
      "grad_norm": 0.1382758468389511,
      "learning_rate": 2e-07,
      "loss": 0.0432,
      "step": 2129
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1987995030012425,
      "grad_norm": 0.1382252722978592,
      "learning_rate": 2e-07,
      "loss": 0.0303,
      "step": 2130
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1988928361012431,
      "grad_norm": 0.13032706081867218,
      "learning_rate": 2e-07,
      "loss": 0.0342,
      "step": 2131
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19898616920124365,
      "grad_norm": 0.13422849774360657,
      "learning_rate": 2e-07,
      "loss": 0.001,
      "step": 2132
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19907950230124424,
      "grad_norm": 0.13972826302051544,
      "learning_rate": 2e-07,
      "loss": 0.008,
      "step": 2133
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19917283540124484,
      "grad_norm": 0.14237481355667114,
      "learning_rate": 2e-07,
      "loss": -0.0133,
      "step": 2134
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1992661685012454,
      "grad_norm": 0.12254750728607178,
      "learning_rate": 2e-07,
      "loss": -0.0098,
      "step": 2135
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.199359501601246,
      "grad_norm": 0.12538908421993256,
      "learning_rate": 2e-07,
      "loss": -0.0114,
      "step": 2136
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19945283470124658,
      "grad_norm": 0.1335274875164032,
      "learning_rate": 2e-07,
      "loss": 0.0143,
      "step": 2137
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19954616780124718,
      "grad_norm": 0.1363525390625,
      "learning_rate": 2e-07,
      "loss": -0.0333,
      "step": 2138
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19963950090124774,
      "grad_norm": 0.12447486072778702,
      "learning_rate": 2e-07,
      "loss": 0.0264,
      "step": 2139
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19973283400124833,
      "grad_norm": 0.1231338232755661,
      "learning_rate": 2e-07,
      "loss": 0.0623,
      "step": 2140
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.19982616710124892,
      "grad_norm": 0.13693514466285706,
      "learning_rate": 2e-07,
      "loss": 0.0119,
      "step": 2141
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.1999195002012495,
      "grad_norm": 0.13250130414962769,
      "learning_rate": 2e-07,
      "loss": -0.0269,
      "step": 2142
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20001283330125008,
      "grad_norm": 0.126338392496109,
      "learning_rate": 2e-07,
      "loss": 0.0172,
      "step": 2143
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20010616640125067,
      "grad_norm": 0.1334684193134308,
      "learning_rate": 2e-07,
      "loss": 0.049,
      "step": 2144
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20019949950125124,
      "grad_norm": 0.16273587942123413,
      "learning_rate": 2e-07,
      "loss": 0.032,
      "step": 2145
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20029283260125183,
      "grad_norm": 0.12837401032447815,
      "learning_rate": 2e-07,
      "loss": -0.0034,
      "step": 2146
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20038616570125242,
      "grad_norm": 0.11919860541820526,
      "learning_rate": 2e-07,
      "loss": 0.0311,
      "step": 2147
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.200479498801253,
      "grad_norm": 0.14974726736545563,
      "learning_rate": 2e-07,
      "loss": -0.0173,
      "step": 2148
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20057283190125358,
      "grad_norm": 0.12926071882247925,
      "learning_rate": 2e-07,
      "loss": 0.0275,
      "step": 2149
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20066616500125417,
      "grad_norm": 0.13605505228042603,
      "learning_rate": 2e-07,
      "loss": 0.0283,
      "step": 2150
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20075949810125474,
      "grad_norm": 0.14126770198345184,
      "learning_rate": 2e-07,
      "loss": -0.0326,
      "step": 2151
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20085283120125533,
      "grad_norm": 0.2894771695137024,
      "learning_rate": 2e-07,
      "loss": 0.0647,
      "step": 2152
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20094616430125592,
      "grad_norm": 0.13805672526359558,
      "learning_rate": 2e-07,
      "loss": 0.0097,
      "step": 2153
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20103949740125648,
      "grad_norm": 0.16674578189849854,
      "learning_rate": 2e-07,
      "loss": -0.0191,
      "step": 2154
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20113283050125708,
      "grad_norm": 0.13608667254447937,
      "learning_rate": 2e-07,
      "loss": 0.0144,
      "step": 2155
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20122616360125767,
      "grad_norm": 0.13364991545677185,
      "learning_rate": 2e-07,
      "loss": 0.0863,
      "step": 2156
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20131949670125826,
      "grad_norm": 0.1404200792312622,
      "learning_rate": 2e-07,
      "loss": 0.005,
      "step": 2157
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20141282980125882,
      "grad_norm": 0.16753411293029785,
      "learning_rate": 2e-07,
      "loss": 0.0109,
      "step": 2158
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20150616290125942,
      "grad_norm": 0.13755065202713013,
      "learning_rate": 2e-07,
      "loss": 0.0564,
      "step": 2159
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20159949600126,
      "grad_norm": 0.15062540769577026,
      "learning_rate": 2e-07,
      "loss": 0.0459,
      "step": 2160
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20169282910126057,
      "grad_norm": 0.13488022983074188,
      "learning_rate": 2e-07,
      "loss": -0.0031,
      "step": 2161
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20178616220126117,
      "grad_norm": 0.14370547235012054,
      "learning_rate": 2e-07,
      "loss": 0.0303,
      "step": 2162
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20187949530126176,
      "grad_norm": 0.14172962307929993,
      "learning_rate": 2e-07,
      "loss": -0.0111,
      "step": 2163
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20197282840126232,
      "grad_norm": 0.13476237654685974,
      "learning_rate": 2e-07,
      "loss": 0.0357,
      "step": 2164
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20206616150126291,
      "grad_norm": 0.1525455266237259,
      "learning_rate": 2e-07,
      "loss": -0.0049,
      "step": 2165
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2021594946012635,
      "grad_norm": 0.13718955218791962,
      "learning_rate": 2e-07,
      "loss": 0.0217,
      "step": 2166
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20225282770126407,
      "grad_norm": 0.1331365704536438,
      "learning_rate": 2e-07,
      "loss": 0.0386,
      "step": 2167
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20234616080126466,
      "grad_norm": 0.1426374316215515,
      "learning_rate": 2e-07,
      "loss": 0.0006,
      "step": 2168
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20243949390126526,
      "grad_norm": 0.1502739042043686,
      "learning_rate": 2e-07,
      "loss": 0.0516,
      "step": 2169
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20253282700126582,
      "grad_norm": 0.14679895341396332,
      "learning_rate": 2e-07,
      "loss": 0.007,
      "step": 2170
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2026261601012664,
      "grad_norm": 0.14002646505832672,
      "learning_rate": 2e-07,
      "loss": 0.0381,
      "step": 2171
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.202719493201267,
      "grad_norm": 0.16434414684772491,
      "learning_rate": 2e-07,
      "loss": 0.0556,
      "step": 2172
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20281282630126757,
      "grad_norm": 0.14029616117477417,
      "learning_rate": 2e-07,
      "loss": -0.0146,
      "step": 2173
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20290615940126816,
      "grad_norm": 0.15086455643177032,
      "learning_rate": 2e-07,
      "loss": 0.06,
      "step": 2174
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20299949250126875,
      "grad_norm": 0.13477592170238495,
      "learning_rate": 2e-07,
      "loss": 0.0468,
      "step": 2175
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20309282560126932,
      "grad_norm": 0.1533563733100891,
      "learning_rate": 2e-07,
      "loss": 0.0095,
      "step": 2176
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.010114397321428603,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4081.0,
      "completions/mean_length": 627.794921875,
      "completions/mean_terminated_length": 592.3577270507812,
      "completions/min_length": 85.0,
      "completions/min_terminated_length": 85.0,
      "epoch": 0.2031861587012699,
      "grad_norm": 0.12680374085903168,
      "learning_rate": 2e-07,
      "loss": 0.0135,
      "num_tokens": 1476804146.0,
      "reward": 0.6364397406578064,
      "reward_std": 0.1716804802417755,
      "rewards/simpleverify_reward/mean": 0.6364397406578064,
      "rewards/simpleverify_reward/std": 0.4810282588005066,
      "step": 2177
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2032794918012705,
      "grad_norm": 0.13784533739089966,
      "learning_rate": 2e-07,
      "loss": 0.0165,
      "step": 2178
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2033728249012711,
      "grad_norm": 0.1357499212026596,
      "learning_rate": 2e-07,
      "loss": 0.0141,
      "step": 2179
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20346615800127166,
      "grad_norm": 0.1367364376783371,
      "learning_rate": 2e-07,
      "loss": 0.0288,
      "step": 2180
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20355949110127225,
      "grad_norm": 0.1486557126045227,
      "learning_rate": 2e-07,
      "loss": 0.0366,
      "step": 2181
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20365282420127284,
      "grad_norm": 0.12914809584617615,
      "learning_rate": 2e-07,
      "loss": 0.0411,
      "step": 2182
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2037461573012734,
      "grad_norm": 0.13629017770290375,
      "learning_rate": 2e-07,
      "loss": 0.029,
      "step": 2183
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.203839490401274,
      "grad_norm": 0.13471633195877075,
      "learning_rate": 2e-07,
      "loss": 0.044,
      "step": 2184
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2039328235012746,
      "grad_norm": 0.13050241768360138,
      "learning_rate": 2e-07,
      "loss": 0.0171,
      "step": 2185
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20402615660127515,
      "grad_norm": 0.13189242780208588,
      "learning_rate": 2e-07,
      "loss": 0.0159,
      "step": 2186
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20411948970127575,
      "grad_norm": 0.13076217472553253,
      "learning_rate": 2e-07,
      "loss": 0.0522,
      "step": 2187
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20421282280127634,
      "grad_norm": 0.1266879290342331,
      "learning_rate": 2e-07,
      "loss": 0.0081,
      "step": 2188
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2043061559012769,
      "grad_norm": 0.15487824380397797,
      "learning_rate": 2e-07,
      "loss": 0.0271,
      "step": 2189
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2043994890012775,
      "grad_norm": 0.1416330635547638,
      "learning_rate": 2e-07,
      "loss": -0.0033,
      "step": 2190
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2044928221012781,
      "grad_norm": 0.14031192660331726,
      "learning_rate": 2e-07,
      "loss": 0.0407,
      "step": 2191
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20458615520127865,
      "grad_norm": 0.14013458788394928,
      "learning_rate": 2e-07,
      "loss": 0.0393,
      "step": 2192
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20467948830127924,
      "grad_norm": 0.13578231632709503,
      "learning_rate": 2e-07,
      "loss": 0.0078,
      "step": 2193
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20477282140127984,
      "grad_norm": 0.12682318687438965,
      "learning_rate": 2e-07,
      "loss": -0.0261,
      "step": 2194
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2048661545012804,
      "grad_norm": 0.12077781558036804,
      "learning_rate": 2e-07,
      "loss": 0.014,
      "step": 2195
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.204959487601281,
      "grad_norm": 0.13998205959796906,
      "learning_rate": 2e-07,
      "loss": 0.0243,
      "step": 2196
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20505282070128159,
      "grad_norm": 0.11716945469379425,
      "learning_rate": 2e-07,
      "loss": 0.037,
      "step": 2197
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20514615380128218,
      "grad_norm": 0.13920149207115173,
      "learning_rate": 2e-07,
      "loss": 0.0311,
      "step": 2198
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20523948690128274,
      "grad_norm": 0.1269838511943817,
      "learning_rate": 2e-07,
      "loss": 0.0311,
      "step": 2199
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20533282000128333,
      "grad_norm": 0.16309025883674622,
      "learning_rate": 2e-07,
      "loss": 0.0078,
      "step": 2200
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20542615310128393,
      "grad_norm": 0.14266115427017212,
      "learning_rate": 2e-07,
      "loss": 0.0378,
      "step": 2201
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2055194862012845,
      "grad_norm": 0.1387840062379837,
      "learning_rate": 2e-07,
      "loss": -0.0118,
      "step": 2202
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20561281930128508,
      "grad_norm": 0.13319367170333862,
      "learning_rate": 2e-07,
      "loss": -0.0146,
      "step": 2203
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20570615240128567,
      "grad_norm": 0.13310669362545013,
      "learning_rate": 2e-07,
      "loss": 0.0308,
      "step": 2204
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20579948550128624,
      "grad_norm": 0.12871991097927094,
      "learning_rate": 2e-07,
      "loss": 0.0469,
      "step": 2205
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20589281860128683,
      "grad_norm": 0.12678378820419312,
      "learning_rate": 2e-07,
      "loss": 0.0002,
      "step": 2206
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20598615170128742,
      "grad_norm": 0.13681796193122864,
      "learning_rate": 2e-07,
      "loss": -0.0137,
      "step": 2207
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.206079484801288,
      "grad_norm": 0.14061179757118225,
      "learning_rate": 2e-07,
      "loss": 0.0623,
      "step": 2208
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20617281790128858,
      "grad_norm": 0.12841878831386566,
      "learning_rate": 2e-07,
      "loss": 0.0103,
      "step": 2209
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20626615100128917,
      "grad_norm": 0.13524489104747772,
      "learning_rate": 2e-07,
      "loss": 0.0338,
      "step": 2210
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20635948410128974,
      "grad_norm": 0.12110421061515808,
      "learning_rate": 2e-07,
      "loss": 0.0493,
      "step": 2211
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20645281720129033,
      "grad_norm": 0.1355554312467575,
      "learning_rate": 2e-07,
      "loss": -0.0267,
      "step": 2212
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20654615030129092,
      "grad_norm": 0.14319929480552673,
      "learning_rate": 2e-07,
      "loss": 0.0102,
      "step": 2213
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20663948340129148,
      "grad_norm": 0.13738730549812317,
      "learning_rate": 2e-07,
      "loss": 0.0249,
      "step": 2214
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20673281650129208,
      "grad_norm": 0.13152654469013214,
      "learning_rate": 2e-07,
      "loss": -0.0043,
      "step": 2215
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20682614960129267,
      "grad_norm": 0.1363997757434845,
      "learning_rate": 2e-07,
      "loss": 0.0259,
      "step": 2216
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20691948270129323,
      "grad_norm": 0.1330546736717224,
      "learning_rate": 2e-07,
      "loss": 0.038,
      "step": 2217
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20701281580129383,
      "grad_norm": 0.12124863266944885,
      "learning_rate": 2e-07,
      "loss": 0.013,
      "step": 2218
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20710614890129442,
      "grad_norm": 0.14265406131744385,
      "learning_rate": 2e-07,
      "loss": 0.0454,
      "step": 2219
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.207199482001295,
      "grad_norm": 0.13326601684093475,
      "learning_rate": 2e-07,
      "loss": 0.045,
      "step": 2220
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20729281510129557,
      "grad_norm": 0.14154662191867828,
      "learning_rate": 2e-07,
      "loss": 0.0147,
      "step": 2221
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20738614820129617,
      "grad_norm": 0.13577722012996674,
      "learning_rate": 2e-07,
      "loss": -0.029,
      "step": 2222
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20747948130129676,
      "grad_norm": 0.13199861347675323,
      "learning_rate": 2e-07,
      "loss": 0.0437,
      "step": 2223
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20757281440129732,
      "grad_norm": 0.13864371180534363,
      "learning_rate": 2e-07,
      "loss": 0.0142,
      "step": 2224
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20766614750129792,
      "grad_norm": 0.13210584223270416,
      "learning_rate": 2e-07,
      "loss": 0.0534,
      "step": 2225
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2077594806012985,
      "grad_norm": 0.12624220550060272,
      "learning_rate": 2e-07,
      "loss": -0.0017,
      "step": 2226
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20785281370129907,
      "grad_norm": 0.17965185642242432,
      "learning_rate": 2e-07,
      "loss": 0.0245,
      "step": 2227
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20794614680129966,
      "grad_norm": 0.137176513671875,
      "learning_rate": 2e-07,
      "loss": 0.0185,
      "step": 2228
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20803947990130026,
      "grad_norm": 0.13284356892108917,
      "learning_rate": 2e-07,
      "loss": 0.0839,
      "step": 2229
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20813281300130082,
      "grad_norm": 0.14883238077163696,
      "learning_rate": 2e-07,
      "loss": -0.0086,
      "step": 2230
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2082261461013014,
      "grad_norm": 0.14229613542556763,
      "learning_rate": 2e-07,
      "loss": 0.0463,
      "step": 2231
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.208319479201302,
      "grad_norm": 0.1320336014032364,
      "learning_rate": 2e-07,
      "loss": 0.0207,
      "step": 2232
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20841281230130257,
      "grad_norm": 0.13614624738693237,
      "learning_rate": 2e-07,
      "loss": 0.0313,
      "step": 2233
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20850614540130316,
      "grad_norm": 0.13710087537765503,
      "learning_rate": 2e-07,
      "loss": 0.0244,
      "step": 2234
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20859947850130375,
      "grad_norm": 0.14568546414375305,
      "learning_rate": 2e-07,
      "loss": 0.0227,
      "step": 2235
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20869281160130432,
      "grad_norm": 0.13461199402809143,
      "learning_rate": 2e-07,
      "loss": 0.0749,
      "step": 2236
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2087861447013049,
      "grad_norm": 0.13689970970153809,
      "learning_rate": 2e-07,
      "loss": -0.0181,
      "step": 2237
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2088794778013055,
      "grad_norm": 0.11833404004573822,
      "learning_rate": 2e-07,
      "loss": 0.0521,
      "step": 2238
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20897281090130607,
      "grad_norm": 0.13790574669837952,
      "learning_rate": 2e-07,
      "loss": 0.0366,
      "step": 2239
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20906614400130666,
      "grad_norm": 0.13585473597049713,
      "learning_rate": 2e-07,
      "loss": 0.0397,
      "step": 2240
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.010585239955357095,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4092.0,
      "completions/mean_length": 643.7006225585938,
      "completions/mean_terminated_length": 606.7662353515625,
      "completions/min_length": 83.0,
      "completions/min_terminated_length": 83.0,
      "epoch": 0.20915947710130725,
      "grad_norm": 0.11351791024208069,
      "learning_rate": 2e-07,
      "loss": 0.0011,
      "num_tokens": 1519406149.0,
      "reward": 0.6213204860687256,
      "reward_std": 0.17462831735610962,
      "rewards/simpleverify_reward/mean": 0.6213204264640808,
      "rewards/simpleverify_reward/std": 0.4850623607635498,
      "step": 2241
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20925281020130784,
      "grad_norm": 0.14349308609962463,
      "learning_rate": 2e-07,
      "loss": 0.0218,
      "step": 2242
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2093461433013084,
      "grad_norm": 0.12408077716827393,
      "learning_rate": 2e-07,
      "loss": 0.0374,
      "step": 2243
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.209439476401309,
      "grad_norm": 0.1251942366361618,
      "learning_rate": 2e-07,
      "loss": -0.0147,
      "step": 2244
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2095328095013096,
      "grad_norm": 0.13669657707214355,
      "learning_rate": 2e-07,
      "loss": -0.0164,
      "step": 2245
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20962614260131016,
      "grad_norm": 0.11843916773796082,
      "learning_rate": 2e-07,
      "loss": 0.0416,
      "step": 2246
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20971947570131075,
      "grad_norm": 0.1312740594148636,
      "learning_rate": 2e-07,
      "loss": 0.0419,
      "step": 2247
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.20981280880131134,
      "grad_norm": 0.12000513076782227,
      "learning_rate": 2e-07,
      "loss": 0.0008,
      "step": 2248
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2099061419013119,
      "grad_norm": 0.1360626369714737,
      "learning_rate": 2e-07,
      "loss": 0.0131,
      "step": 2249
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2099994750013125,
      "grad_norm": 0.12431520223617554,
      "learning_rate": 2e-07,
      "loss": 0.0612,
      "step": 2250
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2100928081013131,
      "grad_norm": 0.12137124687433243,
      "learning_rate": 2e-07,
      "loss": 0.0171,
      "step": 2251
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21018614120131365,
      "grad_norm": 0.1274702101945877,
      "learning_rate": 2e-07,
      "loss": 0.0361,
      "step": 2252
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21027947430131425,
      "grad_norm": 0.15389187633991241,
      "learning_rate": 2e-07,
      "loss": 0.025,
      "step": 2253
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21037280740131484,
      "grad_norm": 0.13659945130348206,
      "learning_rate": 2e-07,
      "loss": 0.0464,
      "step": 2254
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2104661405013154,
      "grad_norm": 0.12139477580785751,
      "learning_rate": 2e-07,
      "loss": 0.0245,
      "step": 2255
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.210559473601316,
      "grad_norm": 0.14551663398742676,
      "learning_rate": 2e-07,
      "loss": 0.0459,
      "step": 2256
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2106528067013166,
      "grad_norm": 0.1292411983013153,
      "learning_rate": 2e-07,
      "loss": -0.0302,
      "step": 2257
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21074613980131715,
      "grad_norm": 0.13494211435317993,
      "learning_rate": 2e-07,
      "loss": 0.0259,
      "step": 2258
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21083947290131774,
      "grad_norm": 0.1279744803905487,
      "learning_rate": 2e-07,
      "loss": 0.0587,
      "step": 2259
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21093280600131833,
      "grad_norm": 0.12434187531471252,
      "learning_rate": 2e-07,
      "loss": -0.0053,
      "step": 2260
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21102613910131893,
      "grad_norm": 0.12350655347108841,
      "learning_rate": 2e-07,
      "loss": 0.0447,
      "step": 2261
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2111194722013195,
      "grad_norm": 0.13516467809677124,
      "learning_rate": 2e-07,
      "loss": 0.0352,
      "step": 2262
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21121280530132008,
      "grad_norm": 0.14506122469902039,
      "learning_rate": 2e-07,
      "loss": -0.0034,
      "step": 2263
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21130613840132068,
      "grad_norm": 0.12468710541725159,
      "learning_rate": 2e-07,
      "loss": 0.0409,
      "step": 2264
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21139947150132124,
      "grad_norm": 0.12582720816135406,
      "learning_rate": 2e-07,
      "loss": 0.003,
      "step": 2265
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21149280460132183,
      "grad_norm": 0.13199901580810547,
      "learning_rate": 2e-07,
      "loss": 0.0292,
      "step": 2266
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21158613770132242,
      "grad_norm": 0.1256171464920044,
      "learning_rate": 2e-07,
      "loss": 0.0249,
      "step": 2267
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.211679470801323,
      "grad_norm": 0.13507325947284698,
      "learning_rate": 2e-07,
      "loss": -0.0002,
      "step": 2268
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21177280390132358,
      "grad_norm": 0.12713368237018585,
      "learning_rate": 2e-07,
      "loss": 0.0476,
      "step": 2269
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21186613700132417,
      "grad_norm": 0.13991907238960266,
      "learning_rate": 2e-07,
      "loss": 0.0596,
      "step": 2270
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21195947010132474,
      "grad_norm": 0.1254131942987442,
      "learning_rate": 2e-07,
      "loss": -0.0131,
      "step": 2271
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21205280320132533,
      "grad_norm": 0.14195223152637482,
      "learning_rate": 2e-07,
      "loss": 0.0292,
      "step": 2272
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21214613630132592,
      "grad_norm": 0.12449552863836288,
      "learning_rate": 2e-07,
      "loss": 0.0115,
      "step": 2273
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21223946940132649,
      "grad_norm": 0.14159028232097626,
      "learning_rate": 2e-07,
      "loss": -0.0045,
      "step": 2274
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21233280250132708,
      "grad_norm": 0.12826327979564667,
      "learning_rate": 2e-07,
      "loss": 0.0026,
      "step": 2275
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21242613560132767,
      "grad_norm": 0.1404125988483429,
      "learning_rate": 2e-07,
      "loss": 0.0625,
      "step": 2276
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21251946870132823,
      "grad_norm": 0.14949820935726166,
      "learning_rate": 2e-07,
      "loss": 0.0298,
      "step": 2277
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21261280180132883,
      "grad_norm": 0.13975505530834198,
      "learning_rate": 2e-07,
      "loss": 0.0067,
      "step": 2278
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21270613490132942,
      "grad_norm": 0.14105868339538574,
      "learning_rate": 2e-07,
      "loss": 0.0092,
      "step": 2279
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21279946800132998,
      "grad_norm": 0.1289086639881134,
      "learning_rate": 2e-07,
      "loss": -0.0123,
      "step": 2280
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21289280110133058,
      "grad_norm": 0.14238393306732178,
      "learning_rate": 2e-07,
      "loss": 0.0284,
      "step": 2281
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21298613420133117,
      "grad_norm": 0.149738147854805,
      "learning_rate": 2e-07,
      "loss": 0.0456,
      "step": 2282
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21307946730133176,
      "grad_norm": 0.1313241720199585,
      "learning_rate": 2e-07,
      "loss": 0.0281,
      "step": 2283
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21317280040133232,
      "grad_norm": 0.14210982620716095,
      "learning_rate": 2e-07,
      "loss": 0.0126,
      "step": 2284
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21326613350133292,
      "grad_norm": 0.12760142982006073,
      "learning_rate": 2e-07,
      "loss": 0.0226,
      "step": 2285
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2133594666013335,
      "grad_norm": 0.13607583940029144,
      "learning_rate": 2e-07,
      "loss": 0.0626,
      "step": 2286
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21345279970133407,
      "grad_norm": 0.1333509385585785,
      "learning_rate": 2e-07,
      "loss": 0.0021,
      "step": 2287
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21354613280133466,
      "grad_norm": 0.13705310225486755,
      "learning_rate": 2e-07,
      "loss": 0.0315,
      "step": 2288
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21363946590133526,
      "grad_norm": 0.15610744059085846,
      "learning_rate": 2e-07,
      "loss": 0.0385,
      "step": 2289
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21373279900133582,
      "grad_norm": 0.13528165221214294,
      "learning_rate": 2e-07,
      "loss": 0.0944,
      "step": 2290
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2138261321013364,
      "grad_norm": 0.15812267363071442,
      "learning_rate": 2e-07,
      "loss": 0.0481,
      "step": 2291
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.213919465201337,
      "grad_norm": 0.1344677209854126,
      "learning_rate": 2e-07,
      "loss": 0.0048,
      "step": 2292
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21401279830133757,
      "grad_norm": 0.13811402022838593,
      "learning_rate": 2e-07,
      "loss": 0.062,
      "step": 2293
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21410613140133816,
      "grad_norm": 0.13607700169086456,
      "learning_rate": 2e-07,
      "loss": 0.0908,
      "step": 2294
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21419946450133875,
      "grad_norm": 0.13305304944515228,
      "learning_rate": 2e-07,
      "loss": -0.0208,
      "step": 2295
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21429279760133932,
      "grad_norm": 0.14102695882320404,
      "learning_rate": 2e-07,
      "loss": 0.0151,
      "step": 2296
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2143861307013399,
      "grad_norm": 0.12146824598312378,
      "learning_rate": 2e-07,
      "loss": 0.0352,
      "step": 2297
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2144794638013405,
      "grad_norm": 0.14284424483776093,
      "learning_rate": 2e-07,
      "loss": 0.0351,
      "step": 2298
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21457279690134107,
      "grad_norm": 0.11069469898939133,
      "learning_rate": 2e-07,
      "loss": 0.0269,
      "step": 2299
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21466613000134166,
      "grad_norm": 0.13185881078243256,
      "learning_rate": 2e-07,
      "loss": 0.0139,
      "step": 2300
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21475946310134225,
      "grad_norm": 0.12587696313858032,
      "learning_rate": 2e-07,
      "loss": 0.0278,
      "step": 2301
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21485279620134284,
      "grad_norm": 0.1323939561843872,
      "learning_rate": 2e-07,
      "loss": 0.0077,
      "step": 2302
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2149461293013434,
      "grad_norm": 0.14714030921459198,
      "learning_rate": 2e-07,
      "loss": 0.0295,
      "step": 2303
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.215039462401344,
      "grad_norm": 0.13663747906684875,
      "learning_rate": 2e-07,
      "loss": -0.0263,
      "step": 2304
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.010271344866071397,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4087.0,
      "completions/mean_length": 632.8683471679688,
      "completions/mean_terminated_length": 596.9281616210938,
      "completions/min_length": 93.0,
      "completions/min_terminated_length": 93.0,
      "epoch": 0.2151327955013446,
      "grad_norm": 0.1281842738389969,
      "learning_rate": 2e-07,
      "loss": 0.0707,
      "num_tokens": 1561440709.0,
      "reward": 0.6358293890953064,
      "reward_std": 0.16936269402503967,
      "rewards/simpleverify_reward/mean": 0.6358293890953064,
      "rewards/simpleverify_reward/std": 0.48120108246803284,
      "step": 2305
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21522612860134516,
      "grad_norm": 0.1206863522529602,
      "learning_rate": 2e-07,
      "loss": 0.028,
      "step": 2306
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21531946170134575,
      "grad_norm": 0.11955222487449646,
      "learning_rate": 2e-07,
      "loss": 0.0428,
      "step": 2307
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21541279480134634,
      "grad_norm": 0.11896443367004395,
      "learning_rate": 2e-07,
      "loss": 0.0444,
      "step": 2308
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2155061279013469,
      "grad_norm": 0.12493612617254257,
      "learning_rate": 2e-07,
      "loss": 0.0201,
      "step": 2309
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2155994610013475,
      "grad_norm": 0.14842964708805084,
      "learning_rate": 2e-07,
      "loss": -0.0116,
      "step": 2310
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2156927941013481,
      "grad_norm": 0.13239946961402893,
      "learning_rate": 2e-07,
      "loss": 0.0182,
      "step": 2311
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21578612720134865,
      "grad_norm": 0.13741400837898254,
      "learning_rate": 2e-07,
      "loss": 0.0348,
      "step": 2312
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21587946030134925,
      "grad_norm": 0.1294623762369156,
      "learning_rate": 2e-07,
      "loss": 0.0509,
      "step": 2313
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21597279340134984,
      "grad_norm": 0.13177455961704254,
      "learning_rate": 2e-07,
      "loss": 0.0228,
      "step": 2314
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2160661265013504,
      "grad_norm": 0.1309087574481964,
      "learning_rate": 2e-07,
      "loss": 0.0447,
      "step": 2315
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.216159459601351,
      "grad_norm": 0.13819552958011627,
      "learning_rate": 2e-07,
      "loss": 0.0275,
      "step": 2316
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2162527927013516,
      "grad_norm": 0.12594059109687805,
      "learning_rate": 2e-07,
      "loss": 0.0309,
      "step": 2317
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21634612580135215,
      "grad_norm": 0.1281934678554535,
      "learning_rate": 2e-07,
      "loss": 0.0194,
      "step": 2318
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21643945890135274,
      "grad_norm": 0.14670638740062714,
      "learning_rate": 2e-07,
      "loss": 0.0018,
      "step": 2319
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21653279200135334,
      "grad_norm": 0.14128999412059784,
      "learning_rate": 2e-07,
      "loss": 0.0563,
      "step": 2320
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2166261251013539,
      "grad_norm": 0.12362868338823318,
      "learning_rate": 2e-07,
      "loss": 0.0145,
      "step": 2321
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2167194582013545,
      "grad_norm": 0.1351068913936615,
      "learning_rate": 2e-07,
      "loss": 0.0204,
      "step": 2322
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21681279130135508,
      "grad_norm": 0.13755571842193604,
      "learning_rate": 2e-07,
      "loss": 0.0185,
      "step": 2323
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21690612440135568,
      "grad_norm": 0.1624317467212677,
      "learning_rate": 2e-07,
      "loss": 0.0205,
      "step": 2324
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21699945750135624,
      "grad_norm": 0.12980917096138,
      "learning_rate": 2e-07,
      "loss": 0.0097,
      "step": 2325
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21709279060135683,
      "grad_norm": 0.14179974794387817,
      "learning_rate": 2e-07,
      "loss": 0.0443,
      "step": 2326
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21718612370135743,
      "grad_norm": 0.11776971817016602,
      "learning_rate": 2e-07,
      "loss": 0.0006,
      "step": 2327
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.217279456801358,
      "grad_norm": 0.13906636834144592,
      "learning_rate": 2e-07,
      "loss": 0.0555,
      "step": 2328
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21737278990135858,
      "grad_norm": 0.1274702548980713,
      "learning_rate": 2e-07,
      "loss": -0.01,
      "step": 2329
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21746612300135917,
      "grad_norm": 0.1376921832561493,
      "learning_rate": 2e-07,
      "loss": 0.0302,
      "step": 2330
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21755945610135974,
      "grad_norm": 0.12755465507507324,
      "learning_rate": 2e-07,
      "loss": 0.0226,
      "step": 2331
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21765278920136033,
      "grad_norm": 0.1292119175195694,
      "learning_rate": 2e-07,
      "loss": 0.0206,
      "step": 2332
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21774612230136092,
      "grad_norm": 0.13331297039985657,
      "learning_rate": 2e-07,
      "loss": 0.0274,
      "step": 2333
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2178394554013615,
      "grad_norm": 0.14038382470607758,
      "learning_rate": 2e-07,
      "loss": 0.0308,
      "step": 2334
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21793278850136208,
      "grad_norm": 0.13483187556266785,
      "learning_rate": 2e-07,
      "loss": 0.0172,
      "step": 2335
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21802612160136267,
      "grad_norm": 0.12990142405033112,
      "learning_rate": 2e-07,
      "loss": 0.0536,
      "step": 2336
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21811945470136324,
      "grad_norm": 0.1341685801744461,
      "learning_rate": 2e-07,
      "loss": 0.0293,
      "step": 2337
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21821278780136383,
      "grad_norm": 0.12747296690940857,
      "learning_rate": 2e-07,
      "loss": 0.0039,
      "step": 2338
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21830612090136442,
      "grad_norm": 0.12237270921468735,
      "learning_rate": 2e-07,
      "loss": 0.0464,
      "step": 2339
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21839945400136498,
      "grad_norm": 0.13124117255210876,
      "learning_rate": 2e-07,
      "loss": 0.0649,
      "step": 2340
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21849278710136558,
      "grad_norm": 0.11371566355228424,
      "learning_rate": 2e-07,
      "loss": -0.0112,
      "step": 2341
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21858612020136617,
      "grad_norm": 0.13129395246505737,
      "learning_rate": 2e-07,
      "loss": 0.044,
      "step": 2342
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21867945330136673,
      "grad_norm": 0.13029909133911133,
      "learning_rate": 2e-07,
      "loss": -0.0044,
      "step": 2343
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21877278640136733,
      "grad_norm": 0.13145920634269714,
      "learning_rate": 2e-07,
      "loss": 0.0218,
      "step": 2344
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21886611950136792,
      "grad_norm": 0.14842307567596436,
      "learning_rate": 2e-07,
      "loss": 0.0169,
      "step": 2345
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2189594526013685,
      "grad_norm": 0.12834519147872925,
      "learning_rate": 2e-07,
      "loss": 0.0503,
      "step": 2346
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21905278570136907,
      "grad_norm": 0.1272391974925995,
      "learning_rate": 2e-07,
      "loss": 0.0131,
      "step": 2347
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21914611880136967,
      "grad_norm": 0.13513964414596558,
      "learning_rate": 2e-07,
      "loss": 0.0132,
      "step": 2348
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21923945190137026,
      "grad_norm": 0.13601651787757874,
      "learning_rate": 2e-07,
      "loss": 0.0034,
      "step": 2349
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21933278500137082,
      "grad_norm": 0.13371379673480988,
      "learning_rate": 2e-07,
      "loss": 0.0125,
      "step": 2350
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21942611810137141,
      "grad_norm": 0.1232561469078064,
      "learning_rate": 2e-07,
      "loss": 0.0378,
      "step": 2351
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.219519451201372,
      "grad_norm": 0.14330853521823883,
      "learning_rate": 2e-07,
      "loss": 0.0338,
      "step": 2352
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21961278430137257,
      "grad_norm": 0.13117177784442902,
      "learning_rate": 2e-07,
      "loss": 0.0047,
      "step": 2353
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21970611740137316,
      "grad_norm": 0.12683221697807312,
      "learning_rate": 2e-07,
      "loss": 0.0099,
      "step": 2354
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21979945050137376,
      "grad_norm": 0.122899129986763,
      "learning_rate": 2e-07,
      "loss": -0.0044,
      "step": 2355
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.21989278360137432,
      "grad_norm": 0.127985417842865,
      "learning_rate": 2e-07,
      "loss": 0.0392,
      "step": 2356
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2199861167013749,
      "grad_norm": 0.12008952349424362,
      "learning_rate": 2e-07,
      "loss": 0.0156,
      "step": 2357
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2200794498013755,
      "grad_norm": 0.13086745142936707,
      "learning_rate": 2e-07,
      "loss": 0.0197,
      "step": 2358
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22017278290137607,
      "grad_norm": 0.1184086874127388,
      "learning_rate": 2e-07,
      "loss": 0.0382,
      "step": 2359
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22026611600137666,
      "grad_norm": 0.13691842555999756,
      "learning_rate": 2e-07,
      "loss": 0.0018,
      "step": 2360
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22035944910137725,
      "grad_norm": 0.13860419392585754,
      "learning_rate": 2e-07,
      "loss": 0.0011,
      "step": 2361
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22045278220137782,
      "grad_norm": 0.1200907751917839,
      "learning_rate": 2e-07,
      "loss": 0.0242,
      "step": 2362
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2205461153013784,
      "grad_norm": 0.13657982647418976,
      "learning_rate": 2e-07,
      "loss": 0.0019,
      "step": 2363
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.220639448401379,
      "grad_norm": 0.12590554356575012,
      "learning_rate": 2e-07,
      "loss": 0.0541,
      "step": 2364
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2207327815013796,
      "grad_norm": 0.13369132578372955,
      "learning_rate": 2e-07,
      "loss": 0.0193,
      "step": 2365
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22082611460138016,
      "grad_norm": 0.12960048019886017,
      "learning_rate": 2e-07,
      "loss": 0.038,
      "step": 2366
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22091944770138075,
      "grad_norm": 0.1389169991016388,
      "learning_rate": 2e-07,
      "loss": -0.0024,
      "step": 2367
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22101278080138134,
      "grad_norm": 0.13383693993091583,
      "learning_rate": 2e-07,
      "loss": 0.0107,
      "step": 2368
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0101318359375,
      "completions/max_length": 4096.0,
      "completions/max_terminated_length": 4090.0,
      "completions/mean_length": 630.2313232421875,
      "completions/mean_terminated_length": 594.75732421875,
      "completions/min_length": 101.0,
      "completions/min_terminated_length": 101.0,
      "epoch": 0.2211061139013819,
      "grad_norm": 0.13836167752742767,
      "learning_rate": 2e-07,
      "loss": 0.0412,
      "num_tokens": 1603306448.0,
      "reward": 0.630685031414032,
      "reward_std": 0.16827768087387085,
      "rewards/simpleverify_reward/mean": 0.6306849718093872,
      "rewards/simpleverify_reward/std": 0.48262354731559753,
      "step": 2369
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2211994470013825,
      "grad_norm": 0.15980982780456543,
      "learning_rate": 2e-07,
      "loss": 0.0297,
      "step": 2370
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2212927801013831,
      "grad_norm": 0.12128932029008865,
      "learning_rate": 2e-07,
      "loss": 0.0037,
      "step": 2371
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22138611320138366,
      "grad_norm": 0.14348870515823364,
      "learning_rate": 2e-07,
      "loss": 0.0195,
      "step": 2372
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22147944630138425,
      "grad_norm": 0.12883640825748444,
      "learning_rate": 2e-07,
      "loss": 0.0048,
      "step": 2373
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22157277940138484,
      "grad_norm": 0.14356324076652527,
      "learning_rate": 2e-07,
      "loss": 0.0414,
      "step": 2374
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2216661125013854,
      "grad_norm": 0.12526337802410126,
      "learning_rate": 2e-07,
      "loss": 0.0135,
      "step": 2375
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.221759445601386,
      "grad_norm": 0.1389387995004654,
      "learning_rate": 2e-07,
      "loss": 0.0172,
      "step": 2376
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2218527787013866,
      "grad_norm": 0.12593252956867218,
      "learning_rate": 2e-07,
      "loss": 0.0213,
      "step": 2377
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22194611180138715,
      "grad_norm": 0.1451130509376526,
      "learning_rate": 2e-07,
      "loss": 0.0589,
      "step": 2378
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22203944490138774,
      "grad_norm": 0.1360453963279724,
      "learning_rate": 2e-07,
      "loss": 0.0411,
      "step": 2379
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22213277800138834,
      "grad_norm": 0.1293872594833374,
      "learning_rate": 2e-07,
      "loss": 0.031,
      "step": 2380
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2222261111013889,
      "grad_norm": 0.14260068535804749,
      "learning_rate": 2e-07,
      "loss": -0.0169,
      "step": 2381
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2223194442013895,
      "grad_norm": 0.13326534628868103,
      "learning_rate": 2e-07,
      "loss": -0.0326,
      "step": 2382
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22241277730139009,
      "grad_norm": 0.14006340503692627,
      "learning_rate": 2e-07,
      "loss": 0.0154,
      "step": 2383
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22250611040139065,
      "grad_norm": 0.12549535930156708,
      "learning_rate": 2e-07,
      "loss": 0.0346,
      "step": 2384
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22259944350139124,
      "grad_norm": 0.13087137043476105,
      "learning_rate": 2e-07,
      "loss": 0.0127,
      "step": 2385
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22269277660139183,
      "grad_norm": 0.1305152177810669,
      "learning_rate": 2e-07,
      "loss": 0.0032,
      "step": 2386
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22278610970139243,
      "grad_norm": 0.13426345586776733,
      "learning_rate": 2e-07,
      "loss": -0.023,
      "step": 2387
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.222879442801393,
      "grad_norm": 0.13531267642974854,
      "learning_rate": 2e-07,
      "loss": 0.0707,
      "step": 2388
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22297277590139358,
      "grad_norm": 0.13917961716651917,
      "learning_rate": 2e-07,
      "loss": 0.0696,
      "step": 2389
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22306610900139417,
      "grad_norm": 0.133058100938797,
      "learning_rate": 2e-07,
      "loss": -0.0011,
      "step": 2390
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22315944210139474,
      "grad_norm": 0.136958047747612,
      "learning_rate": 2e-07,
      "loss": 0.0428,
      "step": 2391
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22325277520139533,
      "grad_norm": 0.1331188976764679,
      "learning_rate": 2e-07,
      "loss": 0.0489,
      "step": 2392
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22334610830139592,
      "grad_norm": 0.12682807445526123,
      "learning_rate": 2e-07,
      "loss": 0.0557,
      "step": 2393
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.2234394414013965,
      "grad_norm": 0.14918924868106842,
      "learning_rate": 2e-07,
      "loss": 0.0415,
      "step": 2394
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22353277450139708,
      "grad_norm": 0.12926289439201355,
      "learning_rate": 2e-07,
      "loss": 0.0046,
      "step": 2395
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22362610760139767,
      "grad_norm": 0.11474764347076416,
      "learning_rate": 2e-07,
      "loss": 0.0149,
      "step": 2396
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22371944070139824,
      "grad_norm": 0.13529391586780548,
      "learning_rate": 2e-07,
      "loss": 0.0207,
      "step": 2397
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22381277380139883,
      "grad_norm": 0.12376414984464645,
      "learning_rate": 2e-07,
      "loss": 0.0359,
      "step": 2398
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22390610690139942,
      "grad_norm": 0.1280665248632431,
      "learning_rate": 2e-07,
      "loss": 0.0509,
      "step": 2399
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "epoch": 0.22399944000139999,
      "grad_norm": 0.1242837980389595,
      "learning_rate": 2e-07,
      "loss": 0.0324,
      "step": 2400
    },
    {
      "epoch": 0.22399944000139999,
      "step": 2400,
      "total_flos": 0.0,
      "train_loss": 0.004557218201868333,
      "train_runtime": 34670.8361,
      "train_samples_per_second": 62.023,
      "train_steps_per_second": 0.069
    }
  ],
  "logging_steps": 1,
  "max_steps": 2400,
  "num_input_tokens_seen": 1603306448,
  "num_train_epochs": 1,
  "save_steps": 160,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}