Natural Language Inference on ANLI (Best-of-K Accuracy)
79.4Best-of-K AccuracyVeriBound
Evaluation Results
| Method | Links | |
|---|---|---|
| VeriBoundK=5, Evaluation Mode=Empirical2026.06 | 79.4 | |
| FOVERK=52026.06 | 78.9 | |
| ReasonFlux-PRMK=52026.06 | 77.8 | |
| VeriBoundK=5, Evaluation Mode=Bound2026.06 | 77.6 | |
| R-PRMK=52026.06 | 77.1 | |
| PRM800KK=52026.06 | 76.3 | |
| Math-ShepherdK=52026.06 | 74.5 | |
| ORMK=52026.06 | 71.2 |