Question Answering on OpenBookQA (Composite Score)
92.14Composite ScoreSFTKey-Tag
Evaluation Results
| Method | Links | |
|---|---|---|
| SFTKey-TagModel=Qwen2.5-7B2025.12 | 92.14 | |
| SFT-TagModel=Qwen3-8B-Base2025.12 | 90.62 | |
| SFT(baseline)Model=Qwen3-8B-Base2025.12 | 90.2 | |
| SFTKey-TagModel=Qwen3-8B-Base2025.12 | 90.2 | |
| SFT-TagModel=Qwen2.5-7B2025.12 | 89.92 | |
| SFT(baseline)Model=Qwen2.5-7B2025.12 | 89.22 | |
| Key-TagModel=Qwen3-8B-Base2025.12 | 87.62 | |
| SFT(baseline)Model=SmolLM3-3B-Base2025.12 | 85.16 | |
| SFTKey-TagModel=SmolLM3-3B-Base2025.12 | 82 | |
| SFT-TagModel=SmolLM3-3B-Base2025.12 | 71.34 | |
| Key-TagModel=Qwen2.5-7B2025.12 | 62.72 | |
| Key-TagModel=SmolLM3-3B-Base2025.12 | 61.92 | |
| Key-TagModel=Qwen2.5-1.5B2025.12 | 58.66 | |
| SFT-TagModel=Qwen2.5-1.5B2025.12 | 57.54 | |
| Key-TagModel=Qwen2.5-3B2025.12 | 55.3 | |
| SFTKey-TagModel=Qwen2.5-1.5B2025.12 | 52.08 | |
| SFT(baseline)Model=Qwen2.5-3B2025.12 | 51.38 | |
| SFT-TagModel=Qwen2.5-3B2025.12 | 49.84 | |
| SFTKey-TagModel=Qwen2.5-3B2025.12 | 49.7 | |
| SFT(baseline)Model=Qwen2.5-1.5B2025.12 | 45.36 |