Selective Generation on GSM8k (PRR)
64.2PRR (Accuracy)SATRMD+MSP
Evaluation Results
| Method | Links | |
|---|---|---|
| SATRMD+MSPModel=Llama 8b v3.12025.02 | 64.2 | |
| SAPLMAModel=Llama 8b v3.12025.02 | 59.8 | |
| HUQ-SATRMDModel=Llama 8b v3.12025.02 | 59.2 | |
| Lexical Similarity ROUGE-LModel=Llama 8b v3.12025.02 | 46.7 | |
| SARModel=Llama 8b v3.12025.02 | 45.5 | |
| EigenScoreModel=Llama 8b v3.12025.02 | 43 | |
| Semantic EntropyModel=Llama 8b v3.12025.02 | 42.4 | |
| Eccentricity NLI Score Entail.Model=Llama 8b v3.12025.02 | 40.3 | |
| Maximum Sequence ProbabilityModel=Llama 8b v3.12025.02 | 38 | |
| DegMat NLI Score Entail.Model=Llama 8b v3.12025.02 | 35.7 | |
| EigValLaplacian NLI Score Entail.Model=Llama 8b v3.12025.02 | 33.5 | |
| PerplexityModel=Llama 8b v3.12025.02 | 25.9 | |
| SentenceSARModel=Llama 8b v3.12025.02 | 15.1 | |
| FactoscopeModel=Llama 8b v3.12025.02 | 4.8 |