Reading Comprehension on AdversarialQA

Metric: D(BERT): F1 (higher is better)

LeaderboardDataset
Loading chart...
#ModelD(BERT): F1Extra DataPaperDateCode
1RoBERTa-Large65.5YesBeat the AI: Investigating Adversarial Human Ann...2020-02-02Code
2BERT-Large62.4YesBeat the AI: Investigating Adversarial Human Ann...2020-02-02Code
3BiDAF30.2YesBeat the AI: Investigating Adversarial Human Ann...2020-02-02Code