| linguistic-acceptability-on-cola-glue-dev | DeBERTa Large | #1 | Matthews Correlation: 70.5 |
| natural-language-inference-on-qnli-glue-dev | DeBERTa Large | #1 | Accuracy: 95.3 |
| natural-language-inference-on-rte-glue-dev | DeBERTa Large | #1 | Accuracy: 88.3 |
| natural-language-inference-on-wnli | DeBERTa | #1 | Accuracy: 94.5 |
| semantic-textual-similarity-on-sts-b-glue-dev | DeBERTa Large | #1 | Pearson/Spearman Correlation: 92.8 |
| common-sense-reasoning-on-record | DeBERTa-1.5B | #2 | EM: 94.1F1: 94.5 |
| math-word-problem-solving-on-paramawps | DeBERTa | #2 | Accuracy (%): 74.1 |
| natural-language-inference-on-multinli-dev | DeBERTa Large | #2 | Matched: 91.1 |
| question-answering-on-quora-question-pairs | DeBERTa (large) | #2 | Accuracy: 92.3% |
| question-answering-on-squad20-dev | DeBERTa Large | #2 | F1: 90.7 |
| linguistic-acceptability-on-cola-dev | DeBERTa (large) | #3 | Accuracy: 69.5 |
| natural-language-inference-on-rte | DeBERTa-1.5B | #3 | Accuracy: 93.2% |
| natural-language-understanding-on-glue-dev | DeBERTa Large | #3 | GLUE (dev): 90.00 |
| question-answering-on-multirc | DeBERTa-1.5B | #3 | F1: 88.2EM: 63.7 |
| coreference-resolution-on-winograd-schema-challenge | DeBERTa-1.5B | #4 | Accuracy: 95.9 |
| natural-language-inference-on-commitmentbank | DeBERTa-1.5B | #4 | Accuracy: 97.2F1: 94.9 |
| question-answering-on-copa | DeBERTa-Ensemble | #4 | Accuracy: 98.4 |
| question-answering-on-copa | DeBERTa-1.5B | #5 | Accuracy: 96.8 |
| word-sense-disambiguation-on-words-in-context | DeBERTa-Ensemble | #4 | Accuracy: 77.5 |
| word-sense-disambiguation-on-words-in-context | DeBERTa-1.5B | #7 | Accuracy: 76.4 |
| natural-language-inference-on-multinli | DeBERTa (large) | #5 | Matched: 91.1Mismatched: 91.1 |
| question-answering-on-boolq | DeBERTa-1.5B | #7 | Accuracy: 90.4 |
| reading-comprehension-on-race | DeBERTalarge | #7 | Accuracy: 86.8 |
| question-answering-on-squad20 | DeBERTalarge | #8 | EM: 88.0F1: 90.7 |
| natural-language-inference-on-qnli | DeBERTa (large) | #10 | Accuracy: 95.3% |
| sentence-completion-on-hellaswag | DeBERTa++ | #13 | Accuracy: 93 |
| sentiment-analysis-on-sst-2-binary-classification | DeBERTa (large) | #15 | Accuracy: 96.5 |
| question-answering-on-squad11-dev | DeBERTa Large | #48 | F1: 95.5 |
| semantic-textual-similarity-on-sts-benchmark | DeBERTa (large) | #57 | Accuracy: 92.5 |