Mathematical Formalization on PutnamBench 672 problems
63C@1Mathesis-7B
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Mathesis-7BTemperature=0.62026.04 | 63 | 88.8 | 43 | 69.3 | 20 | |
| Kimina-7BTemperature=0.62026.04 | 53.5 | 85.7 | 36.8 | 65.3 | 16.7 | |
| SFT+GRPO-0%Temperature=1.02026.04 | 47.9 | 78.1 | 23.6 | 43 | 24.3 | |
| SFT+GRPO-30%Temperature=1.02026.04 | 46.4 | 71.9 | 22.9 | 38.8 | 23.5 | |
| GRPO-onlyTemperature=1.02026.04 | 36.1 | 60.1 | 11.9 | 19.2 | 24.2 | |
| SFT+GRPO-100%Temperature=1.02026.04 | 29.1 | 65.3 | 14.7 | 34.8 | 14.4 | |
| SFTTemperature=1.02026.04 | 28.5 | 65.6 | 14.3 | 34.2 | 14.2 | |
| BaseTemperature=1.02026.04 | 11.3 | 29.3 | 3.3 | 7.7 | 8 |