| question-answering-on-race | Finetuned Transformer LM | #2 | RACE-m: 62.9RACE-h: 57.4RACE: 59.0 |
| natural-language-inference-on-scitail | Finetuned Transformer LM | #3 | Accuracy: 88.3 |
| question-answering-on-storycloze | Finetuned Transformer LM | #7 | Accuracy: 86.5 |
| natural-language-understanding-on-glue-test | GPT-1 | #9 | GLUE (test): 72.8 |
| natural-language-inference-on-snli | Fine-Tuned LM-Pretrained Transformer | #11 | % Test Accuracy: 89.9% Train Accuracy: 96.6Parameters: 85m |
| reading-comprehension-on-race | GPT-1 | #26 | Accuracy: 59.0 |
| paraphrase-identification-on-quora-question-pairs | GPT-1 | #29 | F1: 70.3 |
| natural-language-inference-on-multinli | Finetuned Transformer LM | #34 | Matched: 82.1Mismatched: 81.4 |
| linguistic-acceptability-on-cola | GPT-1 | #36 | Accuracy: 45.4% |
| semantic-textual-similarity-on-mrpc | GPT-1 | #36 | F1: 82.3 |
| natural-language-inference-on-qnli | GPT-1 | #38 | Accuracy: 88.1% |
| sentiment-analysis-on-sst-2-binary-classification | GPT-1 | #53 | Accuracy: 91.3 |
| natural-language-inference-on-rte | GPT-1 | #77 | Accuracy: 56.0% |