Selective Generation on SciQ (PRR (AlignScore))
65.3PRR (AlignScore)HUQ-SATRMD
Evaluation Results
| Method | Links | |
|---|---|---|
| HUQ-SATRMDModel=Llama 8b v3.12025.02 | 65.3 | |
| Maximum Sequence ProbabilityModel=Llama 8b v3.12025.02 | 58.2 | |
| SentenceSARModel=Llama 8b v3.12025.02 | 54.3 | |
| SATRMD+MSPModel=Llama 8b v3.12025.02 | 54.2 | |
| Semantic EntropyModel=Llama 8b v3.12025.02 | 46.6 | |
| DegMat NLI Score Entail.Model=Llama 8b v3.12025.02 | 44.6 | |
| Eccentricity NLI Score Entail.Model=Llama 8b v3.12025.02 | 44.4 | |
| SARModel=Llama 8b v3.12025.02 | 44 | |
| EigValLaplacian NLI Score Entail.Model=Llama 8b v3.12025.02 | 39.8 | |
| SAPLMAModel=Llama 8b v3.12025.02 | 38.8 | |
| EigenScoreModel=Llama 8b v3.12025.02 | 37.3 | |
| Lexical Similarity ROUGE-LModel=Llama 8b v3.12025.02 | 36 | |
| FactoscopeModel=Llama 8b v3.12025.02 | 31.6 | |
| PerplexityModel=Llama 8b v3.12025.02 | 19.7 |