Question Answering on TriviaQA (str-EM @5, @10)
75.97str-EM Accuracy (@5)COCOA
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| COCOAModel=Phi-3-medium (14B)2026.07 | 75.97 | 76.52 | |
| DCCDModel=Phi-3-medium (14B)2026.07 | 75.47 | 77.06 | |
| fullModel=Phi-3-medium (14B)2026.07 | 75.17 | 76.66 | |
| CADModel=Phi-3-medium (14B)2026.07 | 75.06 | 76.6 | |
| DVDModel=Phi-3-medium (14B)2026.07 | 75.06 | 76.1 | |
| ADACADModel=Phi-3-medium (14B)2026.07 | 75.04 | 76.58 | |
| DCCDModel=Qwen3.5-9B2026.07 | 73.08 | 75.1 | |
| fullModel=Qwen3.5-9B2026.07 | 72.47 | 74.3 | |
| ADACADModel=Qwen3.5-9B2026.07 | 72.46 | 74.35 | |
| CADModel=Qwen3.5-9B2026.07 | 72.39 | 74.31 | |
| COCOAModel=Qwen3.5-9B2026.07 | 71.35 | 73.5 | |
| DVDModel=Qwen3.5-9B2026.07 | 71 | 72.44 | |
| DCCDModel=Qwen3.5-2B2026.07 | 70.04 | 72.01 | |
| CADModel=Qwen3.5-2B2026.07 | 69.75 | 71.56 | |
| fullModel=Qwen3.5-2B2026.07 | 69.59 | 71.33 | |
| DVDModel=Qwen3.5-2B2026.07 | 69.57 | 71.13 | |
| ADACADModel=Qwen3.5-2B2026.07 | 69.33 | 71.28 | |
| zero-shotModel=Phi-3-medium (14B)2026.07 | 69.11 | — | |
| COCOAModel=Qwen3.5-2B2026.07 | 66.47 | 67.71 | |
| zero-shotModel=Qwen3.5-9B2026.07 | 57.76 | — | |
| zero-shotModel=Qwen3.5-2B2026.07 | 32.21 | — |