Mathematical Reasoning on OlympiadBench (Pass@8, Pass@16, Pass@32)
63.26Pass@8SENT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SENTBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.12 | 63.26 | 67.26 | 70.22 | |
| w/ EnBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.12 | 60.89 | 62.67 | 69.33 | |
| w/ MaskBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.12 | 60.15 | 65.19 | 69.19 | |
| w/ ClipBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.12 | 58.96 | 64.89 | 68.74 | |
| w/ High EnBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.12 | 58.37 | 63.85 | 66.67 | |
| w/ CovBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.12 | 56.44 | 64.15 | 67.7 | |
| GRPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.12 | 54.96 | 59.41 | 64.3 | |
| w/ AdvBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.12 | 54.81 | 60 | 63.41 | |
| Base modelBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.12 | 52.89 | 58.52 | 61.19 |