| Benchmark | Model | Rank | Results |
|---|---|---|---|
| linguistic-acceptability-on-cola-glue-dev | I-BERT RoBERTa-Large | #4 | Matthews Correlation: 69.0 |
| natural-language-inference-on-rte-glue-dev | I-BERT RoBERTa-Large | #5 | Accuracy: 87.0 |
| natural-language-understanding-on-glue-dev | I-BERT RoBERTa-Large | #5 | GLUE (dev): 89.5 |
| semantic-textual-similarity-on-sts-b-glue-dev | I-BERT RoBERTa-Large | #5 | Pearson/Spearman Correlation: 92.2 |
| natural-language-inference-on-qnli-glue-dev | I-BERT RoBERTa-Large | #8 | Accuracy: 94.5 |