Fact Verification on GSMClaims
61AccuracyThinknCheck-Science-1B
Evaluation results
| Method | Links | |
|---|---|---|
| ThinknCheck-Science-1BModel Size=1B, Training Data=Scientific and arithmetic2026.04 | 61 | |
| ThinknCheck-1BModel Size=1B2026.04 | 52.2 | |
| MiniCheck-7BModel Size=7B2026.04 | 51.3 | |
| ThinknCheck-nothink-1BModel Size=1B, Reasoning=None2026.04 | 49.4 |