Multi-Task Skill Learning Suite (ScienceQA, MMLU, Hellaswag, Humaneval, TruthfulQA, Winogrande, IFeval)
70.2ScienceQASDFT
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| SDFT2026.01 | 70.2 | 60.9 | 68.9 | 66.8 | 70.7 | 46.5 | 73.1 | 64.5 | |
| SFT2026.01 | 66.2 | 55 | 54.8 | 35.3 | 64.6 | 36.8 | 73.7 | 53.4 | |
| SFT + re-invoke2026.01 | 66 | 61.6 | 63.4 | 52.9 | 68.7 | 45.2 | 70 | 60.2 | |
| DFT2026.01 | 54.8 | 57.6 | 67 | 60.4 | 69.4 | 38.8 | 68.2 | 60.2 | |
| BaseBackbone=Qwen2.5-7B2026.01 | 32.1 | 62 | 65.8 | 74.3 | 71.7 | 47.9 | 71.1 | 65.5 |