Factual Consistency Evaluation on SAMSum
46.7Spearman CorrelationALIGN
Evaluation Results
| Method | Links | |
|---|---|---|
| ALIGNModel size=large2023.07 | 46.7 | |
| ALIGNSCORE-largeType=Ours2023.05 | 43.8 | |
| QAFactEvalType=QA2023.05 | 35.9 | |
| QAFactEval2023.07 | 35.9 | |
| MNLIType=NLI2023.05 | 31.1 | |
| ALIGNSCORE-baseType=Ours2023.05 | 23.4 | |
| UniEvalType=Misc2023.05 | 22.8 | |
| UniEval2023.07 | 22.8 | |
| ALIGNModel size=base2023.07 | 22.3 | |
| DAEType=NLI2023.05 | 18.6 | |
| CTCType=Misc2023.05 | 17.7 | |
| CTC2023.07 | 17.7 | |
| ROUGE-2Type=Similarity Matching2023.05 | 17.5 | |
| NER-OverlapType=Similarity Matching2023.05 | 16.7 | |
| ROUGE-LType=Similarity Matching2023.05 | 16.3 | |
| BARTScoreType=Misc2023.05 | 16.3 | |
| BARTScore2023.07 | 16.3 | |
| ROUGE-1Type=Similarity Matching2023.05 | 15.1 | |
| SummaC-ZSType=NLI2023.05 | 15.1 | |
| BLEUType=Similarity Matching2023.05 | 13.7 | |
| BERTScoreType=Similarity Matching2023.05 | 13.1 | |
| BERTScore2023.07 | 13.1 | |
| SummaC-CONVType=NLI2023.05 | 9.8 | |
| BLANCType=Misc2023.05 | 9.1 | |
| SimCSEType=Similarity Matching2023.05 | 7.9 | |
| BLEURTType=Regression2023.05 | 6.7 | |
| BLEURT2023.07 | 6.7 | |
| QuestEvalType=QA2023.05 | 3.9 | |
| FEQAType=QA2023.05 | 0 | |
| FactCCType=Misc2023.05 | -4.4 |