Selective Generation on Aggregate All Datasets
56.3Mean PRRTAD
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TADBackbone=Qwen-2.5 7b2024.08 | 56.3 | 1.27 | |
| LookBackLensBackbone=Qwen-2.5 7b2024.08 | 52.3 | 3.55 | |
| SheepsBackbone=Qwen-2.5 7b2024.08 | 49 | 4.73 | |
| SATRMDBackbone=Qwen-2.5 7b2024.08 | 45.4 | 5.55 | |
| MINDBackbone=Qwen-2.5 7b2024.08 | 44.4 | 7.36 | |
| PerplexityBackbone=Qwen-2.5 7b2024.08 | 33.1 | 10.45 | |
| Mean Token EntropyBackbone=Qwen-2.5 7b2024.08 | 32.2 | 10.55 | |
| Simple FocusBackbone=Qwen-2.5 7b2024.08 | 30.6 | 11.55 | |
| SARBackbone=Qwen-2.5 7b2024.08 | 29.6 | 12.09 | |
| Semantic DensityBackbone=Qwen-2.5 7b2024.08 | 29.6 | 12.27 | |
| LUQBackbone=Qwen-2.5 7b2024.08 | 29.4 | 12.18 | |
| MSPBackbone=Qwen-2.5 7b2024.08 | 29.1 | 12.91 | |
| DegMat NLI Score entail.Backbone=Qwen-2.5 7b2024.08 | 29 | 12.36 | |
| EVL NLI Score entail.Backbone=Qwen-2.5 7b2024.08 | 28.2 | 13.09 | |
| Lexical Similarity Rouge-LBackbone=Qwen-2.5 7b2024.08 | 27.5 | 14 | |
| SAPLMABackbone=Qwen-2.5 7b2024.08 | 27.2 | 14.09 | |
| CCPBackbone=Qwen-2.5 7b2024.08 | 26.4 | 14.36 | |
| FocusBackbone=Qwen-2.5 7b2024.08 | 25.6 | 14.55 | |
| EigenScoreBackbone=Qwen-2.5 7b2024.08 | 23.5 | 16.18 | |
| Ecc. NLI Score entail.Backbone=Qwen-2.5 7b2024.08 | 23.5 | 15.45 | |
| Semantic EntropyBackbone=Qwen-2.5 7b2024.08 | 23.3 | 17.18 | |
| FactoscopeBackbone=Qwen-2.5 7b2024.08 | 22 | 17.27 |