Question Answering on TruthfulQA (PRR Metric)
0.5PRRTAD
Evaluation Results
| Method | Links | |
|---|---|---|
| TADBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.5 | |
| LookBackLensBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.485 | |
| SheepsBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.476 | |
| Mean Token EntropyBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.356 | |
| LUQBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.322 | |
| SAPLMABackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.321 | |
| MINDBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.314 | |
| EVL NLI Score entail.Backbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.306 | |
| DegMat NLI Score entail.Backbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.3 | |
| Semantic DensityBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.3 | |
| PerplexityBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.274 | |
| SATRMDBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.254 | |
| FocusBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.189 | |
| SARBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.185 | |
| Lexical Similarity Rouge-LBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.159 | |
| Ecc. NLI Score entail.Backbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.146 | |
| FactoscopeBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.137 | |
| Simple FocusBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.093 | |
| Semantic EntropyBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | 0.087 | |
| EigenScoreBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | -0.018 | |
| MSPBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | -0.088 | |
| CCPBackbone=Qwen-2.5 7b, Reference Quality Metric=AlignScore2024.08 | -0.104 |