Community Notes Dataset

该数据集包含了66.4万条来自Twitter/X社区的注释，这些注释经过筛选，重点关注其对于错误信息的有用性和相关性。数据集中包含了被评定为“有帮助”、“没有帮助”以及未发布的注释，并且对注释中使用的来源进行了详细的分类。规模上，该数据集共有66.4万条注释。其任务是对与错误信息相关的社区注释进行分析，以及评估这些注释对事实核查来源的依赖程度。

This dataset contains 664,000 annotations sourced from the Twitter/X community, which were filtered to prioritize their usefulness and relevance to misinformation. The dataset includes annotations categorized into three groups: those rated as "helpful", "unhelpful", and unpublished, alongside detailed classification of the sources cited within these annotations. In total, the dataset comprises 664,000 annotations. The core tasks of this dataset involve analyzing community annotations related to misinformation, as well as evaluating the extent to which these annotations rely on fact-checking sources.