Mathematical Reasoning on MATH (test) (Pass@N)
25.34Pass@1W-REINFORCE
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| W-REINFORCEBackbone=Qwen2.5-Math-1.5B2026.05 | 25.34 | 39.85 | 55.28 | 67.98 | 76.82 | 82.88 | 87.62 | 91.54 | 92.1 | |
| CW-NSRBackbone=Qwen2.5-Math-1.5B2026.05 | 22.58 | 34.76 | 47.89 | 59.72 | 69.31 | 76.82 | 82.67 | 87.11 | 90.4 |