Context-aware generation on TriState-Bench Agreement Sagr gold context, prior correct
93.75Exact Match (EM)Greedy
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GreedyModel=Llama2-13B2026.06 | 93.75 | 95.83 | |
| ARRModel=Llama2-13B2026.06 | 93.75 | 95.81 | |
| AdaCADModel=Llama2-13B2026.06 | 93.25 | 95.4 | |
| ARRModel=Qwen2.5-7B2026.06 | 88.25 | 92.27 | |
| GreedyModel=Qwen2.5-7B2026.06 | 86 | 90.76 | |
| CoCoAModel=Llama2-13B2026.06 | 85.5 | 89.46 | |
| COIECDModel=Llama2-13B2026.06 | 84 | 88.25 | |
| AdaCADModel=Mistral-7B2026.06 | 83.25 | 87.95 | |
| GreedyModel=Mistral-7B2026.06 | 82.75 | 87.76 | |
| AdaCADModel=Qwen2.5-7B2026.06 | 82.25 | 88.56 | |
| ARRModel=Mistral-7B2026.06 | 82 | 87.46 | |
| ARRModel=Llama3-8B2026.06 | 76.75 | 83.15 | |
| COIECDModel=Mistral-7B2026.06 | 72.75 | 75.56 | |
| CADModel=Llama2-13B2026.06 | 72.5 | 80.35 | |
| GreedyModel=Llama3-8B2026.06 | 64.75 | 73.92 | |
| COIECDModel=Qwen2.5-7B2026.06 | 56.75 | 71.96 | |
| AdaCADModel=Llama3-8B2026.06 | 48.5 | 60.43 | |
| CADModel=Qwen2.5-7B2026.06 | 36.25 | 56.18 | |
| COIECDModel=Llama3-8B2026.06 | 35 | 50.99 | |
| CoCoAModel=Mistral-7B2026.06 | 32 | 48.19 | |
| CoCoAModel=Qwen2.5-7B2026.06 | 30.75 | 50.6 | |
| CoCoAModel=Llama3-8B2026.06 | 30 | 45.52 | |
| CADModel=Mistral-7B2026.06 | 25 | 27.51 | |
| CADModel=Llama3-8B2026.06 | 21.25 | 37.92 |