NLI4CT

该数据集包含2400条陈述和临床试验报告（CTRs），这些内容经过注释，用于确定推理关系和检索支持性事实。此外，数据集还包括一个竞赛排行榜、官方网站以及复现基线实验所需的代码。规模上，数据集涵盖了2400条陈述和CTRs。任务方面，主要包括两项：一是二元分类（蕴含或矛盾），二是证据选择（排名问题）。

This dataset contains 2400 annotated statements and Clinical Trial Reports (CTRs), which are used to determine inferential relationships and retrieve supporting facts. Additionally, the dataset includes a competition leaderboard, an official website, and the code required to reproduce baseline experiments. In terms of scale, the dataset covers 2400 statements and CTRs. The tasks primarily include two categories: binary classification (entailment or contradiction) and evidence selection (ranking task).