HateBench: Benchmarking Hate Speech Detectors on LLM-Generated Content and Hate Campaigns

To Appear in the USENIX Security 2025
Xinyue Shen1, Yixin Wu1, Yiting Qu1, Michael Backes1, Savvas Zannettou2, Yang Zhang1
1CISPA Helmholtz Center for Information Security, 2TU Delft
PDF arXiv Code 🤗 Data

LLM-Generated Samples

The entire dataset can be accessed at HateBenchSet - Hugging Face.

Disclaimer. This website contains examples of hateful and abusive language. Reader discretion is recommended.


Ethics and Disclosures

Our work relies on LLMs to generate samples, and all the manual annotations are performed by the authors of this study. Therefore our study is not considered human subjects research by our Institutional Review Board (IRB). Also, by doing annotations ourselves, we ensure that no human subjects were exposed to harmful information during our study. Since our work involves the assessment of LLM-driven hate campaigns, it is inevitable to disclose how attackers can evade a hate speech detector. We have taken great care to responsibly share our findings. We disclosed the paper and the labeled dataset to OpenAI, Google Jigsaw, and the developers of open-source detectors. In our disclosure letter, we explicitly highlighted the high attack success rates in the LLM-driven hate campaigns. We have received the acknowledgment from OpenAI and Google Jigsaw.

Website adapted from Nerfies.


BibTeX

If you find this useful in your research, please consider citing:

@inproceedings{SWQBZZ25,
  author = {Xinyue Shen and Yixin Wu and Yiting Qu and Michael Backes and Savvas Zannettou and Yang Zhang},
  title = {{HateBench: Benchmarking Hate Speech Detectors on LLM-Generated Content and Hate Campaigns}},
  booktitle = {{USENIX Security Symposium (USENIX Security)}},
  publisher = {USENIX},
  year = {2025}
}