Overconfident Errors Need Stronger Correction: Asymmetric Confidence Penalties for Reinforcement Learning
Open paper
Benchmark
Model
Rank
Results
mathematical-reasoning-on-aime-2025
Qwen2.5-Math-7B + ACE-DAPO
–
Accuracy: 11.7
mathematical-reasoning-on-math-500
Qwen2.5-Math-7B + ACE-DAPO
–
Accuracy: 75.1
Rank counts only results with a code link.