Factuality Evaluation on QAGS XSUM
53.9Pearson CorrelationALIGN
Evaluation Results
| Method | Links | |
|---|---|---|
| ALIGNModel size=large2023.07 | 53.9 | |
| ALIGNModel size=base2023.07 | 51.1 | |
| UniEval2023.07 | 45.5 | |
| QAFactEval2023.07 | 44.2 | |
| CTC2023.07 | 30.6 | |
| BARTScore2023.07 | 16.3 | |
| BLEURT2023.07 | 13.2 | |
| BARTSCOREFine-tuning task=CNN2021.06 | 0.184 | |
| QAGS2021.06 | 0.175 | |
| ROUGE-22021.06 | 0.097 | |
| BARTSCOREFine-tuning task=CNN + Para2021.06 | 0.074 | |
| MoverScore2021.06 | 0.054 | |
| PRISM2021.06 | 0.025 | |
| ROUGE-L2021.06 | 0.024 | |
| BERTScore2021.06 | 0.024 | |
| BARTSCORESetup=Vanilla2021.06 | 0.009 | |
| ROUGE-12021.06 | -0.008 | |
| BARTSCOREFine-tuning task=CNN, Prompting=true2021.06 | -0.094 | |
| BERTScore2023.07 | -10.6 |