SaFeR-ToolKit (+ SFT+GRPO) [3B]
| Method | Links | |||
|---|---|---|---|---|
| 93.37 | 79.42 | 92.69 | ||
| 93.22 | 89.66 | 94.92 | ||
| 92.88 | 82.2 | 89.49 | ||
2026.06 | 90.49 | 93.32 | - | |
2026.03 | 87.46 | 58.81 | 78.64 | |
2026.03 | 87.44 | 70.46 | 39.05 | |
2026.03 | 84.58 | 53.22 | 50 | |
2026.06 | 84.21 | 75.27 | - | |
2026.06 | 84.04 | 60.75 | - | |
2026.06 | 83.53 | 92.91 | - | |
2026.03 | 83.22 | 70.17 | 83.22 | |
| 82.88 | 75.08 | 82.37 | ||
2026.03 | 79.66 | 8.81 | 17.8 | |
2026.06 | 77.76 | 81.29 | - | |
2026.06 | 76.23 | 82.15 | - | |
2026.06 | 74.53 | 83.94 | - | |
2026.06 | 65.87 | 25.43 | - | |
2026.03 | 63.9 | 74.24 | 26.61 | |
2026.06 | 63.84 | 50.7 | - | |
2026.06 | 62.14 | 62.65 | - | |
2026.06 | 59.59 | 56.45 | - | |
2026.06 | 52.29 | 46.85 | - | |
2026.06 | 50 | 50 | - | |
2026.06 | 50 | 50 | - | |
2026.03 | 37.8 | 65.76 | 6.78 | |
2026.03 | 36.61 | 56.27 | 6.27 | |
2026.03 | 31.02 | 39.66 | 4.75 |