Context-aware generation on TriState-Bench Correction gold context, prior incorrect (Scor)
91.25Exact Match (EM)Greedy
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GreedyModel=Mistral-7B2026.06 | 91.25 | 94.75 | |
| AdaCADModel=Mistral-7B2026.06 | 91 | 94.25 | |
| GreedyModel=Qwen2.5-7B2026.06 | 89.5 | 93.32 | |
| ARRModel=Mistral-7B2026.06 | 88 | 92.73 | |
| GreedyModel=Llama2-13B2026.06 | 86.5 | 90.79 | |
| ARRModel=Qwen2.5-7B2026.06 | 85.75 | 91.09 | |
| AdaCADModel=Llama2-13B2026.06 | 85.25 | 89.83 | |
| COIECDModel=Mistral-7B2026.06 | 85 | 89.46 | |
| ARRModel=Llama2-13B2026.06 | 84.25 | 89.16 | |
| AdaCADModel=Qwen2.5-7B2026.06 | 81.25 | 87.5 | |
| CoCoAModel=Llama2-13B2026.06 | 80.25 | 85.44 | |
| COIECDModel=Llama2-13B2026.06 | 80 | 84.91 | |
| ARRModel=Llama3-8B2026.06 | 75 | 82.68 | |
| CADModel=Mistral-7B2026.06 | 70.5 | 78.48 | |
| CADModel=Llama2-13B2026.06 | 66 | 75.74 | |
| COIECDModel=Qwen2.5-7B2026.06 | 65.75 | 76.63 | |
| GreedyModel=Llama3-8B2026.06 | 59.5 | 70.79 | |
| CoCoAModel=Mistral-7B2026.06 | 55.5 | 70.1 | |
| AdaCADModel=Llama3-8B2026.06 | 51.5 | 64.19 | |
| CoCoAModel=Llama3-8B2026.06 | 31.25 | 47.88 | |
| COIECDModel=Llama3-8B2026.06 | 30.25 | 47.35 | |
| CoCoAModel=Qwen2.5-7B2026.06 | 26.5 | 46.41 | |
| CADModel=Qwen2.5-7B2026.06 | 25.75 | 49.65 | |
| CADModel=Llama3-8B2026.06 | 21.75 | 41.07 |