The annotator agreement and the model performance for individual hate speech classes.
收藏数据链接：
官方服务：
资源简介：
The identification of individual classes is measured by the F1 score. The lines correspond to Table 3. The last three columns give the F1 scores for the three detailed hate speech classes which are merged into a more abstract, Unacceptable class (F1(U)), used throughout the paper. Note relatively low model performance for the Violent class (F1(V)).
创建时间：
2022-03-17
