ai2_arc allenai
Dataset Card for "ai2_arc" Dataset Summary A new dataset of 7,787 genuine grade-school level, multiple-choice science questions, assembled to encourage research in advanced question-answering. The dataset is partitioned into a Challenge Set and an Easy Set, where the former contains only questions answered incorrectly by both a retrieval-based algorithm and a word co-occurrence algorithm. We are also including a corpus of over 14 million science sentences… See the full description on the dataset page: https://huggingface.co/datasets/allenai/ai2_arc.
- 类型
- dataset
- 许可
- cc-by-sa-4.0
- 语言
- en
- 下载量
- 480,331
- 点赞
- 382
- 访问
- public
- 文件
- 0