CoPE is a drop-in enhancement of RoPE that delivers consistent gains within the training context and during long-context extrapoaltion.
Haoran Li PRO
haoranli-ml
·
AI & ML interests
ML, RL, Foundation Models
Recent Activity
published
a dataset
about 14 hours ago
haoranli-ml/genvf-v3-difficulty
updated
a dataset
about 16 hours ago
haoranli-ml/genvf-v3-difficulty
updated
a dataset
3 days ago
haoranli-ml/genvf-v3-train-holdout-60-with_branch_rewards
Organizations
models
17
haoranli-ml/Llama-3-8B-HardClip-64k-Instruct
8B
•
Updated
•
45
haoranli-ml/Llama-3-8B-HardClip-64k-Base
8B
•
Updated
•
36
haoranli-ml/Llama-3-8B-RoPE-64k-Base
8B
•
Updated
•
22
haoranli-ml/Llama-3-8B-CoPE-64k-Base
8B
•
Updated
•
22
haoranli-ml/Llama-3-8B-CoPE-64k-Instruct
Text Generation
•
8B
•
Updated
•
47
haoranli-ml/Llama-3-8B-RoPE-64k-Instruct
Text Generation
•
8B
•
Updated
•
52
haoranli-ml/Qwen3-4B-Instruct_haoran_VF_rl_half_oracle_raw-prefix
4B
•
Updated
haoranli-ml/Qwen3-4B-Instruct_haoran_VF_rl_all_oracle_sum-prefix
4B
•
Updated
haoranli-ml/Qwen3-4B-Instruct_haoran_VF_rl_all_oracle_raw-prefix
4B
•
Updated
haoranli-ml/Qwen3-4B-Instruct_VF_rl_sum-prefix
4B
•
Updated
datasets
95
haoranli-ml/genvf-v3-difficulty
Updated
haoranli-ml/genvf-v3-train-holdout-60-both_prefix
Viewer
•
Updated
•
60
•
54
haoranli-ml/genvf-v3-train-holdout-60-with_branch_rewards
Viewer
•
Updated
•
60
•
7
haoranli-ml/hold_out_60
Viewer
•
Updated
•
60
•
8
haoranli-ml/APEX-Int
Viewer
•
Updated
•
49
•
11
haoranli-ml/AMO-Bench-Int
Viewer
•
Updated
•
30
•
15
haoranli-ml/VF_gemini_summaries_final
Viewer
•
Updated
•
6.89k
•
23
haoranli-ml/VF_gemini_summaries_train_sanity_check_v2
Viewer
•
Updated
•
10
•
11
haoranli-ml/VF_gemini_summaries_train_sanity_check
Viewer
•
Updated
•
10
•
11
haoranli-ml/sanity_check_abstraction_12k_suffix_final
Viewer
•
Updated
•
100
•
18