Refusal on Irrelevant Queries on QASPER
90.68Refusal F1Doc2Atom
Evaluation Results
| Method | Links | |
|---|---|---|
| Doc2AtomBase Model=Qwen3-4B-Instruct2026.06 | 90.68 | |
| Doc2AtomBase Model=Gemma-2-2B-It2026.06 | 58.05 | |
| D2LrawBase Model=Qwen3-4B-Instruct2026.06 | 54.93 | |
| D2LckptBase Model=Qwen3-4B-Instruct2026.06 | 43.06 | |
| D2LatomBase Model=Qwen3-4B-Instruct2026.06 | 39.59 | |
| D2LatomBase Model=Gemma-2-2B-It2026.06 | 25.67 | |
| D2LrawBase Model=Gemma-2-2B-It2026.06 | 23.49 | |
| D2LckptBase Model=Gemma-2-2B-It2026.06 | 1.2 |