Stacey M. Frumm

OpenMTB-Audit: Exposing Over-Refusal and Clinical Expert Perspectives in LLM-Based Molecular Tumor Board Safety Evaluation

This work introduces OpenMTB-Audit, an open-source benchmark of 500 synthetic non-small cell lung cancer cases spanning five adversarial error categories and four safety labels, and developed MTB-AuditAgent, a deterministic seven-module framework separating evidence verification, missing-information detection, safety c...