Claim Verification on SciFact (dev)
64.7BAccThinknCheck
Evaluation Results
| Method | Links | |
|---|---|---|
| ThinknCheckModel Size=1B, Precision=fp42026.04 | 64.7 | |
| MiniCheckModel Size=7B, Precision=fp162026.04 | 50 | |
| ThinknCheck-nothinkModel Size=1B, Precision=fp42026.04 | 21.7 |
Loading the SOTA2 catalog…