Math Word Problem Reasoning on SVAMP (AUROC)
65.83AUROC ScoreProbas-mean + CoT-UQ
Evaluation Results
| Method | Links | |
|---|---|---|
| Probas-mean + CoT-UQStrategy=AP, Backbone=Mistral 7B2025.02 | 65.83 | |
| TOKENSAR + CoT-UQStrategy=AP, Backbone=Mistral 7B2025.02 | 65.42 | |
| TOKENSARStrategy=AP, Backbone=Mistral 7B2025.02 | 60.69 | |
| Probas-meanStrategy=AP, Backbone=Mistral 7B2025.02 | 59.88 | |
| P(True) + CoT-UQStrategy=SE, Backbone=Mistral 7B2025.02 | 50.37 | |
| P(True)Strategy=SE, Backbone=Mistral 7B2025.02 | 42.93 |