Claim Verification on 9 In-Domain Datasets Suite
74.1FEVER AccuracyDECOMPOSERL
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| DECOMPOSERLModel Scale=7B2026.05 | 74.1 | 98.6 | 76.4 | 93.1 | 86.5 | 87.6 | 87.5 | 85.5 | 87.7 | 86.3 | |
| SimpleModel Scale=7B2026.05 | 72.7 | 94.9 | 71 | 93.5 | 83.2 | 82.7 | 84.2 | 84.1 | 86.6 | 83.7 | |
| SimpleModel Scale=3B2026.05 | 71.5 | 94 | 63.7 | 89 | 73.7 | 82 | 81.2 | 78.7 | 79.3 | 79.2 | |
| Decomposed PromptingModel Scale=14B2026.05 | 71.1 | 100 | 75 | 90.9 | 89 | 83.4 | 86.7 | 85.3 | 88.3 | 85.5 | |
| Self-AskModel Scale=GPT-4.1-mini2026.05 | 70.9 | 100 | 76.7 | 93.5 | 87.2 | 88.3 | 86.4 | 87.1 | 91.1 | 86.8 | |
| Decomposed PromptingModel Scale=32B2026.05 | 68.6 | 100 | 76.2 | 93.2 | 91.3 | 85.1 | 86.8 | 87.4 | 90.3 | 86.5 |