| Benchmark | Model | Rank | Results |
|---|---|---|---|
| language-modelling-on-wikitext-103 | GPT-Neo 2.7B | #5 | Test perplexity: 11.39 |
| question-answering-on-pubmedqa | GPT-Neo 2.7B | #31 | Accuracy: 57.54 |
| language-modelling-on-lambada | GPT-Neo 2.7B | #34 | Accuracy: 62.22Perplexity: 5.626 |
| common-sense-reasoning-on-winogrande | GPT-Neo 2.7B | #140 | Accuracy: 56.50 |
| question-answering-on-piqa | GPT-Neo 2.7B | #141 | Accuracy: 72.14 |
| sentence-completion-on-hellaswag | GPT-Neo 2.7B | #194 | Accuracy: 42.73 |