Fact-Checking on LLM-AGGREFACT (test)
0.2Cost ($)AlignScore
Evaluation results
| Method | Links | |
|---|---|---|
| AlignScoreBackbone Model=ROBERTa-L, Model Size=355M, # FT Data=4,700K2024.04 | 0.2 | |
| MiniCheck-RBTABackbone Model=AlignScore, Model Size=355M, # FT Data=14K2024.04 | 0.2 | |
| MiniCheck-DBTABackbone Model=DeBERTa-L, Model Size=355M, # FT Data=35K2024.04 | 0.2 | |
| FT5-ANLI-LBackbone Model=Flan-T5-L, Model Size=770M, # FT Data=163K2024.04 | 0.24 | |
| MiniCheck-FT5Backbone Model=Flan-T5-L, Model Size=770M, # FT Data=35K2024.04 | 0.24 | |
| DAEBackbone Model=ELECTRA-B, Model Size=110M, # FT Data=95K2024.04 | 0.26 | |
| SummaC-ZSBackbone Model=ALBERT-XL, Model Size=60M, # FT Data=371K2024.04 | 0.85 | |
| SummaC-CVBackbone Model=ALBERT-XL, Model Size=60M, # FT Data=381K2024.04 | 0.85 | |
| QAFactEvalBackbone Model=multiple*, Model Size=1.4B2024.04 | 1.87 | |
| T5-NLI-MixedBackbone Model=T5-XXL, Model Size=11B, # FT Data=1,697K2024.04 | 7.39 |