Mathematical Reasoning on AMC23 (Pass@k Suite)
13.34Pass@1 AccuracyA-NSR
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| A-NSRBackbone=Qwen2.5-Math-1.5B2026.05 | 13.34 | 21.87 | 32.17 | 43.1 | 54.22 | 64.29 | 72.55 | 78.76 | 82.5 | |
| CW-NSRBackbone=Qwen2.5-Math-1.5B2026.05 | 10.32 | 17.64 | 27.23 | 37.73 | 48.45 | 58.75 | 67.21 | 72.86 | 75 | |
| W-REINFORCEBackbone=Qwen2.5-Math-1.5B2026.05 | 10.25 | 17.57 | 27.27 | 37.96 | 48.46 | 57.92 | 65.1 | 69.34 | 72.5 |