Mathematical Reasoning on GSM8K (Selection Accuracy)
33.07Selection AccuracySelf-Certainty (SC)
Evaluation results
| Method | Links | |
|---|---|---|
| Self-Certainty (SC)Perturbation=none2026.01 | 33.07 | |
| Contrastive Causality (CC)Perturbation=none2026.01 | 32.97 | |
| Self-Certainty (SC)Perturbation=shuffling2026.01 | 32.23 | |
| Self-Certainty (SC)Perturbation=small-eval2026.01 | 32.17 | |
| Contrastive Causality (CC)Perturbation=small-eval2026.01 | 32.13 | |
| Contrastive Causality (CC)Perturbation=shuffling2026.01 | 31.73 |