Hate speech detection models are only as good as the data they are trained
on. Datasets sourced from social media suffer from systematic gaps and biases,
leading to unreliable models with simplistic decision boundaries. Adversarial
datasets, collected by exploiting model weaknesses, promise to fix this
problem. However, adversarial data collection can be slow and costly, and
individual annotators have limited creativity. In this paper, we introduce
GAHD, a new German Adversarial Hate speech Dataset comprising ca.\ 11k
examples. During data collection, we explore new strategies for supporting
annotators, to create more diverse adversarial examples more efficiently and
provide a manual analysis of annotator disagreements for each strategy. Our
experiments show that the resulting dataset is challenging even for
state-of-the-art hate speech detection models, and that training on GAHD
clearly improves model robustness. Further, we find that mixing multiple
support strategies is most advantageous. We make GAHD publicly available at
this https URL

通过 GAHD 数据集，作者通过新策略支持标记人员创造更多样化的对抗性样本，提高现有仇恨言论检测模型的鲁棒性。