Reading Comprehension on AdversarialQA
Metric: D(RoBERTa): F1 (higher is better)
Metric: D(RoBERTa): F1 (higher is better)
| # | Model↕ | D(RoBERTa): F1▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | BERT-Large | 54.4 | Yes | Beat the AI: Investigating Adversarial Human Ann... | 2020-02-02 | Code |
| 2 | RoBERTa-Large | 53.4 | Yes | Beat the AI: Investigating Adversarial Human Ann... | 2020-02-02 | Code |
| 3 | BiDAF | 26.7 | Yes | Beat the AI: Investigating Adversarial Human Ann... | 2020-02-02 | Code |