Domain-Aware Reinforcement Fine-Tuning
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| 32.3 | 28.8 | 38.9 | 26.4 | 44.4 | 19.1 | 36.4 | ||
2026.01 | 28.5 | 23.6 | 32.9 | 25.6 | 40.2 | 18.2 | 30.5 | |
2026.01 | 25.5 | 21.8 | 29.9 | 22.6 | 39.5 | 15.4 | 23.6 | |
| 25 | 20.1 | 24.4 | 25.8 | 38.8 | 17.5 | 23.5 | ||
2026.01 | 21.9 | 18.4 | 22.4 | 23.5 | 32.6 | 12.5 | 21.8 | |
| 20.9 | 17.1 | 23.8 | 19.6 | 33.6 | 12.6 | 18.7 | ||
2026.01 | 20.3 | 17 | 21.8 | 19.8 | 32.3 | 12 | 18.9 | |
| 19.7 | 16.1 | 22 | 17.1 | 33.7 | 10.8 | 18.4 | ||
2026.01 | 19.1 | 15.7 | 21.8 | 16.2 | 31.7 | 11.9 | 17.2 | |
2026.01 | 18.3 | 14.9 | 20.6 | 16 | 31.2 | 10.4 | 16.6 | |
2026.01 | 17.4 | 10.3 | 20.9 | 16.1 | 30.7 | 9.61 | 16.5 | |
2026.01 | 16.8 | 10.1 | 19.8 | 15.7 | 30.3 | 9.15 | 15.6 | |
2026.01 | 16 | 8.91 | 18.6 | 15.5 | 29.3 | 8.61 | 14.9 |