Refusal on Irrelevant Queries on SQuAD
85.25Refusal F1Doc2Atom
Evaluation Results
| Method | Links | |
|---|---|---|
| Doc2AtomBase Model=Gemma-2-2B-It2026.06 | 85.25 | |
| Doc2AtomBase Model=Qwen3-4B-Instruct2026.06 | 54.13 | |
| D2LckptBase Model=Qwen3-4B-Instruct2026.06 | 42.6 | |
| D2LckptBase Model=Gemma-2-2B-It2026.06 | 7.61 | |
| D2LrawBase Model=Qwen3-4B-Instruct2026.06 | 2.83 | |
| D2LrawBase Model=Gemma-2-2B-It2026.06 | 1.52 | |
| D2LatomBase Model=Gemma-2-2B-It2026.06 | 0.87 | |
| D2LatomBase Model=Qwen3-4B-Instruct2026.06 | 0.65 |