TrustAIRLab/HateBench

[USENIX'25] HateBench: Benchmarking Hate Speech Detectors on LLM-Generated Content and Hate Campaigns

35
/ 100
Emerging

This project helps researchers evaluate how well existing hate speech detection tools can identify harmful content created by large language models (LLMs). It provides a curated dataset of LLM-generated text, labeled as either 'hate' or 'non-hate', including examples crafted to evade detection. Researchers and trust and safety teams can use this to assess detector performance against sophisticated, AI-driven hate campaigns.

No commits in the last 6 months.

Use this if you need to rigorously test and compare the effectiveness of different hate speech detectors, especially against content generated by advanced AI models.

Not ideal if you are looking for an off-the-shelf hate speech detection solution for real-time content moderation or if you are not comfortable working with potentially offensive content for research.

AI-safety content-moderation-research hate-speech-detection LLM-vulnerabilities trust-and-safety
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 5 / 25
Maturity 16 / 25
Community 14 / 25

How are scores calculated?

Stars

13

Forks

3

Language

License

Apache-2.0

Last pushed

Mar 01, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/TrustAIRLab/HateBench"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.