Question Answering on Adversarial SQuAD (AddSent)
75.5EMALUM-RoBERTa-Large
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ALUM-RoBERTa-LargeBackbone=RoBERTa-Large, Training=Adversarial pre-training2020.04 | 75.5 | 69.4 | |
| RoBERTa-LargeBackbone=RoBERTa-Large2020.04 | 72.3 | 66 | |
| ALUM-BERT-BaseBackbone=BERT-Base, Training=Adversarial pre-training2020.04 | 54.6 | 60.4 | |
| BERT+-BaseBackbone=BERT-Base, Training=Adversarial fine-tuning2020.04 | 50.1 | 56.2 | |
| BERT-BaseBackbone=BERT-Base2020.04 | 48.9 | 54 |