CoVERT

该数据集包含了源自推特的健康相关声明，所有声明都具有因果性质，这带来了非正式语言处理的挑战。它针对现实世界中的错误信息核查场景进行了探讨。该数据集的规模为，子集包含264项声明，其中198项得到支持，66项被反驳。任务主要集中在社交媒体上的错误信息检测。

This dataset contains health-related claims sourced from Twitter, all of which are causal in nature, presenting challenges for informal language processing. It is designed to explore real-world misinformation verification scenarios. The dataset has a subset with 264 claims, of which 198 are supported and 66 are refuted. The core task focuses on misinformation detection on social media.