ESPO: Early-Stopping Proximal Policy Optimization
Open paper
Benchmark
Model
Rank
Results
mathematical-reasoning-on-aime24
ESPO (DeepSeek-R1-Distill-Qwen-7B)
–
Acc: 46.28
mathematical-reasoning-on-math-500
ESPO (DeepSeek-R1-Distill-Qwen-7B)
–
Accuracy: 87.42
Rank counts only results with a code link.