Turn-level correlation evaluation on PersonaChat turn-level
0.4814Spearman CorrelationUSR-DR
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| USR-DRMetric Category=Reference Free Metrics, config=x = c2020.05 | 0.4814 | 0.6087 | |
| USRMetric Category=Reference Free Metrics2020.05 | 0.4693 | 0.4115 | |
| METEORMetric Category=Word-Overlap Metrics2020.05 | 0.2527 | 0.2713 | |
| USR2020.05 | 0.243 | 0.1862 | |
| USR-DRx=c2020.05 | 0.2291 | 0.1733 | |
| BERTScoreMetric Category=Embedding Based Metrics, variant=base2020.05 | 0.169 | 0.1526 | |
| Vector ExtremaMetric Category=Embedding Based Metrics2020.05 | 0.157 | 0.141 | |
| BERTScoreMetric Category=Embedding Based Metrics, variant=large2020.05 | 0.1518 | 0.1211 | |
| F-1Metric Category=Word-Overlap Metrics2020.05 | 0.1422 | 0.1241 | |
| VectorExtrema2020.05 | 0.1375 | 0.1458 | |
| BLEU-4Metric Category=Word-Overlap Metrics2020.05 | 0.1353 | 0.0899 | |
| BLEU-3Metric Category=Word-Overlap Metrics2020.05 | 0.1202 | 0.0924 | |
| Embedding AverageMetric Category=Embedding Based Metrics2020.05 | 0.1182 | 0.1428 | |
| BLEU-2Metric Category=Word-Overlap Metrics2020.05 | 0.1122 | 0.0943 | |
| Greedy MatchingMetric Category=Embedding Based Metrics2020.05 | 0.0916 | 0.0625 | |
| USR-MLMMetric Category=Reference Free Metrics2020.05 | 0.0795 | 0.0788 | |
| ROUGE-LMetric Category=Word-Overlap Metrics2020.05 | 0.0659 | 0.0385 | |
| BLEU-1Metric Category=Word-Overlap Metrics2020.05 | 0.0434 | 0.0469 | |
| Skip-ThoughtMetric Category=Embedding Based Metrics2020.05 | -0.0393 | -0.0452 | |
| USR-DRMetric Category=Reference Free Metrics, config=x = f2020.05 | -0.0495 | -0.0454 |