| Benchmark | Model | Rank | Results |
|---|---|---|---|
| linguistic-acceptability-on-cola-glue-dev | RECAP (BERT-base, 50% pruning) | #10 | Matthews Correlation: 57.67 |
| natural-language-inference-on-rte-glue-dev | RECAP (BERT-base, 50% pruning) | #12 | Accuracy: 65.98 |
| paraphrase-identification-on-mrpc-glue-dev | RECAP (BERT-base, 50% pruning) | #12 | F1: 86.55 |
| natural-language-inference-on-qnli-glue-dev | RECAP (BERT-base, 50% pruning) | #13 | Accuracy: 90.74 |