Math Reasoning on GSM MC
3.75FPRRenewal process-based stopping
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Renewal process-based stoppingStopping Rule=Renewal2026.02 | 3.75 | — | — | 69.75 | |
| LengthStopping Rule=Length2026.02 | 4.5 | — | — | 32.17 | |
| EntropyStopping Rule=Entropy2026.02 | 4.83 | — | — | 44.92 | |
| DEERStopping Rule=DEER2026.02 | 5.08 | — | — | 37.25 | |
| Maxwise Conformal StoppingStopping Rule=Maxwise2026.02 | 5.42 | — | — | 70.92 | |
| ConfidenceStopping Rule=Confidence2026.02 | — | 0.42 | 5.42 | — | |
| CriticismStopping Rule=Criticism2026.02 | — | 0.67 | 5.67 | — | |
| No InterventionStopping Rule=No Intervention2026.02 | — | — | 6.75 | — |