
摘要
请提供需要翻译的英文内容。
代码仓库
allenai/rainbow
官方
tf
GitHub 中提及
基准测试
| 基准 | 方法 | 指标 |
|---|---|---|
| common-sense-reasoning-on-commonsenseqa | Unicorn 11B (fine-tuned) | Accuracy: 79.3 |
| common-sense-reasoning-on-winogrande | Unicorn 11B (fine-tuned) | Accuracy: 91.3 |
| question-answering-on-piqa | Unicorn 11B (fine-tuned) | Accuracy: 90.1 |
| question-answering-on-social-iqa | Unicorn 11B (fine-tuned) | Accuracy: 83.2 |