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

查看完整页面 · 查看原文