Standalone benchmark for multi-turn safety persistence in medical LLM conversations. Measures recommendation monotonicity under sustained patient pressure.
-
Updated
Sep 3, 2026 - Python
Standalone benchmark for multi-turn safety persistence in medical LLM conversations. Measures recommendation monotonicity under sustained patient pressure.
To associate your repository with the safety-persistence topic, visit your repo's landing page and select "manage topics."