Multi-answer Machine Reading Comprehension on Quoref (val)
71EM PrecisionTAGGING
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| TAGGINGBackbone=RoBERTa2023.06 | 71 | 72.21 | 71.6 | 80.44 | 79.74 | 80.09 | |
| ITERATIVEBackbone=RoBERTa2023.06 | 67.28 | 66.35 | 66.81 | 78.57 | 78.58 | 78.57 | |
| NUMPREDBackbone=RoBERTa2023.06 | 65.61 | 63.57 | 64.57 | 77.3 | 78.2 | 77.75 | |
| GENERATIONBackbone=BART-base2023.06 | 63.57 | 63.39 | 63.48 | 73.38 | 74.02 | 73.7 |