Accelerating RL for LLM Reasoning with Optimal Advantage Regression
Paper • 2505.20686 • Published • 3
Accelerating RL for LLM Reasoning with Optimal Advantage Reg
-
-
Cornell-AGI/gsm8k_size_qwen2.5_1.5b_eval
Viewer • Updated • 7.47k • 22 -
Cornell-AGI/gsm8k_size_qwen2.5_3b_eval
Viewer • Updated • 7.47k • 28 -
Cornell-AGI/gsm8k_size_qwen2.5_7b_eval
Viewer • Updated • 7.47k • 35 -
Cornell-AGI/math_size_qwen2.5_1.5b_eval
Viewer • Updated • 7.5k • 21 -
Cornell-AGI/math_size_qwen2.5_3b_eval
Viewer • Updated • 7.5k • 33 -
Cornell-AGI/math_size_qwen2.5_7b_eval
Viewer • Updated • 7.5k • 17 -
Cornell-AGI/apo_math_qwen2.5_1.5b
Text Generation • 2B • Updated • 13 -
Cornell-AGI/apo_math_qwen2.5_3b
Text Generation • 3B • Updated • 18 -
Cornell-AGI/apo_math_qwen2.5_7b
Text Generation • 8B • Updated • 17 • 1 -
Cornell-AGI/ppo_math_qwen2.5_1.5b
Text Generation • 2B • Updated • 10 -
Cornell-AGI/ppo_math_qwen2.5_3b
Text Generation • 3B • Updated • 16 -
Cornell-AGI/ppo_math_qwen2.5_7b
Text Generation • 8B • Updated • 11 -
Cornell-AGI/grpo_math_qwen2.5_1.5b
Text Generation • 2B • Updated • 13 -
Cornell-AGI/grpo_math_qwen2.5_3b
Text Generation • 3B • Updated • 22 -
Cornell-AGI/grpo_math_qwen2.5_7b
Text Generation • 8B • Updated • 32 -
Cornell-AGI/rebel_math_qwen2.5_1.5b
Text Generation • 2B • Updated • 11 -
Cornell-AGI/rebel_math_qwen2.5_3b
Text Generation • 3B • Updated • 11 -
Cornell-AGI/rebel_math_qwen2.5_7b
Text Generation • 8B • Updated • 14