aestrad7/llm-break-bench

Quantitative benchmark for llm safety filter effectiveness using MLCommons AI Safety taxonomy

4 best points · 6 comments · 1 Hacker News post · 8 GitHub stars · View on GitHub

Hacker News discussions

• I ran 3,360 safety tests on GPT-4o, Claude, Grok, DeepSeek, Gemini (4 points, 6 comments, by aestrad7, 2026-03-25)