Knowledge-conflict Question Answering on NQ Swap
89.63CF ScoreFIDES
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FIDESBackbone=Qwen1.5-7B2026.06 | 89.63 | 49.82 | 55.27 | |
| FIDESBackbone=LLaMA3-8B2026.06 | 88.23 | 43.43 | 53.84 | |
| CLEAR†Backbone=Qwen1.5-7B2026.06 | 86.41 | 44.18 | 49.53 | |
| AdaCADBackbone=Qwen1.5-7B2026.06 | 85.27 | 41.73 | 46.49 | |
| AdaCADBackbone=LLaMA3-8B2026.06 | 84.23 | 38.12 | 43.35 | |
| COIECDBackbone=Qwen1.5-7B2026.06 | 83.67 | 37.41 | 41.7 | |
| FIDESBackbone=Mistral-7B2026.06 | 83.56 | 43.78 | 48.16 | |
| COIECDBackbone=LLaMA3-8B2026.06 | 82.77 | 34.6 | 38.94 | |
| CLEAR†Backbone=LLaMA3-8B2026.06 | 82.51 | 42.17 | 52.16 | |
| CADBackbone=Qwen1.5-7B2026.06 | 82.37 | 32.12 | 35.84 | |
| FIDESBackbone=LLaMA2-7B2026.06 | 81.67 | 43.16 | 46.24 | |
| CADBackbone=LLaMA3-8B2026.06 | 81.57 | 30.29 | 33.56 | |
| DVDBackbone=Qwen1.5-7B2026.06 | 81.27 | 40.83 | 45.51 | |
| AdaCADBackbone=Mistral-7B2026.06 | 80.54 | 37.04 | 41.96 | |
| CLEAR†Backbone=Mistral-7B2026.06 | 80.42 | 41.23 | 44.82 | |
| COIECDBackbone=Mistral-7B2026.06 | 79.43 | 33.61 | 37.55 | |
| DeCoReBackbone=Qwen1.5-7B2026.06 | 79.12 | 39.71 | 44.16 | |
| CLEAR†Backbone=LLaMA2-7B2026.06 | 78.53 | 40.82 | 43.51 | |
| CADBackbone=Mistral-7B2026.06 | 78.53 | 29.42 | 32.16 | |
| DoLaBackbone=Qwen1.5-7B2026.06 | 78.51 | 38.27 | 41.62 | |
| AdaCADBackbone=LLaMA2-7B2026.06 | 77.94 | 36.43 | 40.71 | |
| DVDBackbone=LLaMA3-8B2026.06 | 76.89 | 37.64 | 42.52 | |
| COIECDBackbone=LLaMA2-7B2026.06 | 76.58 | 32.92 | 36.4 | |
| CADBackbone=LLaMA2-7B2026.06 | 75.46 | 28.64 | 31.13 | |
| Standard RAGBackbone=Qwen1.5-7B2026.06 | 75.43 | 36.82 | 39.51 | |
| DeCoReBackbone=LLaMA3-8B2026.06 | 75.38 | 36.82 | 40.13 | |
| DVDBackbone=Mistral-7B2026.06 | 75.31 | 36.51 | 41.27 | |
| DVDBackbone=LLaMA2-7B2026.06 | 73.28 | 35.94 | 40.09 | |
| DeCoReBackbone=Mistral-7B2026.06 | 72.18 | 35.62 | 39.43 | |
| DoLaBackbone=LLaMA3-8B2026.06 | 71.42 | 35.16 | 38.81 | |
| DeCoReBackbone=LLaMA2-7B2026.06 | 70.37 | 34.81 | 38.86 | |
| Standard RAGBackbone=LLaMA3-8B2026.06 | 70.21 | 34.52 | 37.53 | |
| DoLaBackbone=Mistral-7B2026.06 | 70.16 | 34.29 | 37.81 | |
| DoLaBackbone=LLaMA2-7B2026.06 | 69.14 | 33.22 | 36.87 | |
| Standard RAGBackbone=Mistral-7B2026.06 | 68.42 | 33.17 | 36.42 | |
| Standard RAGBackbone=LLaMA2-7B2026.06 | 66.83 | 32.12 | 35.41 |