e-SNLI
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/oanamariacamburu/e-snli
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是为自然语言推理而设计的，包含了569,000个由（输入、标签、解释）组成的实例，其中这些解释是通过众包方式获得的。该数据集包含三种标签：蕴含、中性以及矛盾。规模上，该数据集共有569,000个实例，任务专注于自然语言推理。
This dataset is designed for natural language inference (NLI) tasks, containing 569,000 instances structured as (input, label, explanation), where the explanations were obtained via crowdsourcing. It includes three standard labels: Entailment, Neutral, and Contradiction. With a total of 569,000 instances, this dataset focuses on natural language inference tasks.
提供机构：
Stanford University
