General Knowledge QA on CSQA
84.1AccuracyDistillable
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DistillableBase Model=Qwen3-8B2026.04 | 84.1 | 84.7 | 91.4 | |
| UndistillableBase Model=Qwen3-8B2026.04 | 83.9 | 84.1 | 92 | |
| TeacherBase Model=Qwen3-8B2026.04 | 83.7 | 83.9 | 90.7 | |
| DistillableBase Model=Gemma-3-12B2026.04 | 80.4 | 80.7 | 89.9 | |
| ImproveBase Model=Qwen3-1.7B2026.04 | 79.2 | 78.1 | 89.2 | |
| UndistillableBase Model=Gemma-3-12B2026.04 | 79.1 | 79.1 | 91.6 | |
| TeacherBase Model=Gemma-3-12B2026.04 | 78.6 | 78.9 | 95 | |
| SFTBase Model=Qwen3-1.7B2026.04 | 74.8 | 75.3 | 93.4 | |
| StudentBase Model=Qwen3-1.7B2026.04 | 74.4 | 74.8 | 88 | |
| GKD-FKLBase Model=Qwen3-1.7B2026.04 | 74.3 | 73.9 | 91.2 | |
| GKD-RKLBase Model=Qwen3-1.7B2026.04 | 74 | 75.2 | 94.7 | |
| ImproveBase Model=Gemma-3-4B2026.04 | 71 | 70.8 | 88.8 | |
| SFTBase Model=Gemma-3-4B2026.04 | 70.8 | 71 | 92 | |
| GKD-FKLBase Model=Gemma-3-4B2026.04 | 70.7 | 71.2 | 91.8 | |
| GKD-RKLBase Model=Gemma-3-4B2026.04 | 70.5 | 71 | 91.5 | |
| StudentBase Model=Gemma-3-4B2026.04 | 68.9 | 68.8 | 87.7 | |
| MisledBase Model=Gemma-3-4B2026.04 | 27.6 | 28.6 | 45 | |
| MisledBase Model=Qwen3-1.7B2026.04 | 26.3 | 32.2 | 30.4 |