Measuring Hate Speech

该数据集是一个广泛且公开可获取的资源，涵盖了来自8,472名人类标注者为39,565条独特的社交媒体帖子分配的135,556个仇恨言论标签。这些帖子是从三个主要在线平台——推特、Reddit和YouTube收集而来的。数据集还包括了关于标注者和目标对象的丰富社会人口信息，这为深入研究社会人口特征如何影响对仇恨言论感知提供了可能。该数据集的规模涉及135,556个标签和39,565条帖子，其任务旨在进行仇恨言论检测以及标注者偏见分析。

This dataset is a broad, publicly available resource containing 135,556 hate speech labels assigned by 8,472 human annotators to 39,565 unique social media posts. These posts were collected from three major online platforms: Twitter, Reddit, and YouTube. The dataset also includes rich sociodemographic information about both annotators and their target subjects, enabling in-depth research on how sociodemographic characteristics influence perceptions of hate speech. Boasting a scale of 135,556 labels and 39,565 posts, this dataset supports core tasks including hate speech detection and annotator bias analysis.