Question Answering on Distractor hard
62.2Accuracy (Distractor hard)NWCAD
Evaluation Results
| Method | Links | |
|---|---|---|
| NWCADModel=Llama-3.1-70B2026.04 | 62.2 | |
| BaselineModel=Llama-3.1-70B2026.04 | 58.4 | |
| With-contextModel=Llama-3.1-70B2026.04 | 57.8 | |
| NWCADModel=Ministral-3-8B2026.04 | 55.83 | |
| AdaCADModel=Llama-3.1-70B2026.04 | 54.8 | |
| AdaCADModel=Ministral-3-8B2026.04 | 53.21 | |
| With-contextModel=Ministral-3-8B2026.04 | 52.82 | |
| NWCADModel=Llama-3.1-8B2026.04 | 52.16 | |
| CoCoAModel=Llama-3.1-70B2026.04 | 52 | |
| With-contextModel=Llama-3.1-8B2026.04 | 51.25 | |
| AdaCADModel=Llama-3.1-8B2026.04 | 48.49 | |
| BaselineModel=Ministral-3-8B2026.04 | 41.42 | |
| CADModel=Ministral-3-8B2026.04 | 41.02 | |
| BaselineModel=Llama-3.1-8B2026.04 | 40.81 | |
| CoCoAModel=Ministral-3-8B2026.04 | 40.5 | |
| CoCoAModel=Llama-3.1-8B2026.04 | 39.71 | |
| CADModel=Llama-3.1-8B2026.04 | 32.5 | |
| CADModel=Llama-3.1-70B2026.04 | 29.8 |