tasksource/measuring-hate-speech-votes

measuring-hate-speech-votes Measuring Hate Speech (Kennedy et al., 2020; Sachdeva et al., 2022), one row per comment with vote counts. The source has one row per (comment, annotator). Each survey item becomes a list of vote counts over its ordinal codes, in code order: 0-4 for the nine Likert items, where a higher code is more hateful (sentiment: strongly positive to strongly negative; respect: strongly respectful to strongly disrespectful; insult, humiliate, dehumanize… See the full description on the dataset page: https://huggingface.co/datasets/tasksource/measuring-hate-speech-votes.