Mathematical Reasoning on Minerva (Avg@32, Pass@32)
39.4Average Score @32SAPO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SAPOBackbone=DeepSeek-R1-Distill-Qwen-7B2026.02 | 39.4 | 58.1 | |
| MASPOBackbone=DeepSeek-R1-Distill-Qwen-7B2026.02 | 39.3 | 58.6 | |
| Clip HigherBackbone=DeepSeek-R1-Distill-Qwen-7B2026.02 | 38.8 | 58.5 | |
| BAPOBackbone=DeepSeek-R1-Distill-Qwen-7B2026.02 | 38.3 | 57.5 | |
| GRPOBackbone=DeepSeek-R1-Distill-Qwen-7B2026.02 | 37.4 | 57.2 | |
| DACBackbone=DeepSeek-R1-Distill-Qwen-7B2026.02 | 36.6 | 58.7 | |
| Entropy Adv.Backbone=DeepSeek-R1-Distill-Qwen-7B2026.02 | 36.5 | 54 | |
| MASPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.02 | 30.7 | 54.1 | |
| SAPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.02 | 29.6 | 52.7 | |
| DACBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.02 | 27.4 | 54 | |
| GRPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.02 | 26.1 | 48.8 | |
| Entropy Adv.Backbone=DeepSeek-R1-Distill-Qwen-1.5B2026.02 | 25.2 | 48.6 | |
| BAPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.02 | 25.1 | 44.7 | |
| Clip HigherBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.02 | 24.8 | 49.3 |