Factual Consistency Evaluation on QAGS CNNDM
73.9Spearman CorrelationALIGNSCORE-large
Evaluation Results
| Method | Links | |
|---|---|---|
| ALIGNSCORE-largeBackbone=ROBERTa (355M)2023.05 | 73.9 | |
| ALIGNSCORE-largeType=Ours2023.05 | 73.9 | |
| ALIGN-largeModel Parameters=355M2023.07 | 71.6 | |
| ALIGNModel size=large2023.07 | 71.6 | |
| ALIGNSCORE-baseBackbone=ROBERTa (125M)2023.05 | 69 | |
| ALIGNSCORE-baseType=Ours2023.05 | 69 | |
| G-EVAL-4Backbone=GPT42023.05 | 68.5 | |
| G-EVAL-4Model ID=GPT42023.07 | 68.5 | |
| BARTScoreType=Misc2023.05 | 68.1 | |
| BARTScore2023.07 | 68.1 | |
| UniEvalType=Misc2023.05 | 67.6 | |
| UniEval2023.07 | 67.6 | |
| QAFactEvalType=QA2023.05 | 63.1 | |
| QAFactEval2023.07 | 63.1 | |
| SummaC-CONVType=NLI2023.05 | 58.4 | |
| CTCType=Misc2023.05 | 57.3 | |
| CTC2023.07 | 57.3 | |
| ALIGN-baseModel Parameters=125M2023.07 | 56.1 | |
| ALIGNModel size=base2023.07 | 56.1 | |
| BLEUType=Similarity Matching2023.05 | 55.4 | |
| ROUGE-2Type=Similarity Matching2023.05 | 55.2 | |
| ROUGE-LType=Similarity Matching2023.05 | 53.7 | |
| ROUGE-1Type=Similarity Matching2023.05 | 53.6 | |
| G-EVAL-3.5Backbone=GPT3.5-d032023.05 | 51.6 | |
| G-EVAL-3.5Model ID=GPT3.5-d032023.07 | 51.6 | |
| SummaC-ZSType=NLI2023.05 | 51.1 | |
| SimCSEType=Similarity Matching2023.05 | 47.2 | |
| BERTScoreType=Similarity Matching2023.05 | 46.2 | |
| BERTScore2023.07 | 46.2 | |
| BLEURTType=Regression2023.05 | 43.4 | |
| BLEURT2023.07 | 43.4 | |
| FactCCType=Misc2023.05 | 40.3 | |
| DAEType=NLI2023.05 | 37.1 | |
| QuestEvalType=QA2023.05 | 30.8 | |
| BLANCType=Misc2023.05 | 22.2 | |
| NER-OverlapType=Similarity Matching2023.05 | 0.2 | |
| FEQAType=QA2023.05 | -7.2 | |
| MNLIType=NLI2023.05 | -16.4 |