Selective Generation on TruthfulQA
0.744ROC-AUCTAD
Evaluation Results
| Method | Links | |
|---|---|---|
| TADBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.744 | |
| LookBackLensBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.739 | |
| TADBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.738 | |
| SheepsBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.736 | |
| MINDBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.729 | |
| TADBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.726 | |
| SAPLMABackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.725 | |
| LookBackLensBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.718 | |
| LookBackLensBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.716 | |
| SheepsBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.703 | |
| MINDBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.696 | |
| SATRMDBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.694 | |
| SheepsBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.687 | |
| SAPLMABackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.68 | |
| MINDBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.68 | |
| Semantic DensityBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.68 | |
| SATRMDBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.679 | |
| LUQBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.663 | |
| LUQBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.657 | |
| Mean Token EntropyBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.652 | |
| EVL NLI Score entail.Backbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.648 | |
| Mean Token EntropyBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.648 | |
| DegMat NLI Score entail.Backbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.642 | |
| PerplexityBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.641 | |
| FocusBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.641 | |
| Semantic DensityBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.634 | |
| SATRMDBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.622 | |
| Mean Token EntropyBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.622 | |
| SARBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.621 | |
| FocusBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.619 | |
| PerplexityBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.617 | |
| PerplexityBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.616 | |
| SARBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.615 | |
| EVL NLI Score entail.Backbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.614 | |
| Simple FocusBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.613 | |
| MSPBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.612 | |
| Semantic DensityBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.612 | |
| DegMat NLI Score entail.Backbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.608 | |
| DegMat NLI Score entail.Base Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.606 | |
| EVL NLI Score entail.Base Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.605 | |
| SAPLMABase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.603 | |
| FocusBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.597 | |
| Simple FocusBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.588 | |
| Lexical Similarity Rouge-LBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.583 | |
| LUQBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.58 | |
| Ecc. NLI Score entail.Backbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.579 | |
| Simple FocusBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.565 | |
| Ecc. NLI Score entail.Backbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.565 | |
| SARBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.564 | |
| Ecc. NLI Score entail.Base Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.559 | |
| Lexical Similarity Rouge-LBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.557 | |
| CCPBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.556 | |
| MSPBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.55 | |
| EigenScoreBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.55 | |
| FactoscopeBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.546 | |
| FactoscopeBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.541 | |
| Semantic EntropyBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.538 | |
| FactoscopeBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.537 | |
| EigenScoreBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.537 | |
| Semantic EntropyBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.529 | |
| Lexical Similarity Rouge-LBase Model=Llama-3.1 8b, Ground Truth Metric=AlignScore2024.08 | 0.525 | |
| CCPBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.52 | |
| EigenScoreBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.502 | |
| Semantic EntropyBackbone=Gemma-2 9b, Reference Metric=AlignScore2024.08 | 0.496 | |
| MSPBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.463 | |
| CCPBackbone=Qwen-2.5 7b, UQ Level=Sequence-level, Reference Metric=AlignScore2024.08 | 0.452 |