-
TinyV: Reducing False Negatives in Verification Improves RL for LLM Reasoning
Paper • 2505.14625 • Published • 13 -
1
TinyV
💬Verify model answers against ground truth
-
zhangchenxu/TinyV-Qwen3-1.7B
Text Generation • 2B • Updated • 5 -
zhangchenxu/TinyV-Qwen3-1.7B-Think
Text Generation • 2B • Updated • 11 • 3
Zhangchen Xu PRO
zhangchenxu
AI & ML interests
LLM Data, Alignment, Post-Training, Safety
Recent Activity
upvoted
a
paper
2 days ago
Building a Foundational Guardrail for General Agentic Systems via
Synthetic Data
new activity
3 days ago
Agent-Ark/Toucan-1.5M:Clarification on SFT dataset construction for reproducing results
liked
a dataset
5 days ago
bigcode/autocodearena-v0
Organizations
-
TinyV: Reducing False Negatives in Verification Improves RL for LLM Reasoning
Paper • 2505.14625 • Published • 13 -
1
TinyV
💬Verify model answers against ground truth
-
zhangchenxu/TinyV-Qwen3-1.7B
Text Generation • 2B • Updated • 5 -
zhangchenxu/TinyV-Qwen3-1.7B-Think
Text Generation • 2B • Updated • 11 • 3
models
40
zhangchenxu/RB-Qwen2.5-VL-3B-Instruct-vlr_syn_filtered_10k_exp6_hybrid_nothink-GRPO_step312
4B
•
Updated
•
1
zhangchenxu/RB-Qwen2.5-VL-3B-Instruct-vlr_syn_filtered_10k_exp6_hybrid_nothink-GRPO_step288
4B
•
Updated
•
4
zhangchenxu/RB-Qwen2.5-VL-3B-Instruct-vlr_syn_filtered_10k_exp6_hybrid_nothink-GRPO_step256
4B
•
Updated
•
5
zhangchenxu/RB-Qwen2.5-VL-3B-Instruct-vlr_syn_filtered_10k_exp6_hybrid_nothink-GRPO_step224
4B
•
Updated
•
4
zhangchenxu/RB-Qwen2.5-VL-3B-Instruct-vlr_syn_filtered_10k_exp6_hybrid_nothink-GRPO_step192
4B
•
Updated
•
5
zhangchenxu/RB-Qwen2.5-VL-3B-Instruct-vlr_syn_filtered_10k_exp6_hybrid_nothink-GRPO_step160
4B
•
Updated
•
3
zhangchenxu/RB-Qwen2.5-VL-3B-Instruct-vlr_syn_filtered_10k_exp6_hybrid_nothink-GRPO_step128
4B
•
Updated
•
4
zhangchenxu/RB-Qwen2.5-VL-3B-Instruct-vlr_syn_filtered_10k_exp6_hybrid_nothink-GRPO_step96
4B
•
Updated
•
1
zhangchenxu/RB-Qwen2.5-VL-3B-Instruct-vlr_syn_filtered_10k_exp6_hybrid_nothink-GRPO_step64
4B
•
Updated
•
1
zhangchenxu/deepseek-math-7b-instruct-deepscaler_5k_prime_step468
7B
•
Updated
•
5
datasets
14
zhangchenxu/HardVerify-Math
Viewer
•
Updated
•
250
•
39
•
1
zhangchenxu/TinyV_Think_Training_Data_Qwen3_Balanced
Viewer
•
Updated
•
113k
•
14
zhangchenxu/TinyV_Training_Data_Qwen3_Balanced
Viewer
•
Updated
•
113k
•
6
zhangchenxu/bigmath_tinyv_filtered
Viewer
•
Updated
•
7.01k
•
21
zhangchenxu/TinyV_Training_Data_Balanced
Viewer
•
Updated
•
159k
•
4
zhangchenxu/TinyV_Think_Training_Data_Balanced
Viewer
•
Updated
•
159k
•
5
zhangchenxu/KodCode_50K_R1
Viewer
•
Updated
•
50k
•
6
zhangchenxu/KodCode_Hard_18K_R1
Viewer
•
Updated
•
18.7k
•
6
zhangchenxu/Magpie-100k-Gemma2-9B
Viewer
•
Updated
•
100k
•
9
zhangchenxu/zero-eval
Viewer
•
Updated
•
800
•
323