Tencent-Hunyuan/DiagSafe
DiagSafe: diagnostic LLM safety benchmark with prompt-level rubrics (3,489 bilingual prompts, 84-label taxonomy)
GitHub repository with 31 stars and 2 forks.
Language: Python
Topics: benchmark, bilingual, dataset, jailbreak, llm-as-a-judge, llm-evaluation, llm-safety, red-teaming, rubric, safety-benchmark