SOURCE-LINKED INTELLIGENCE

Right Diagnoses, Decorative Reasoning:A Perturbation Audit of Medical Chain-of-Thought

arXiv · AI, language, vision and robotics · article · Aug 25, 2026 · UTC

Clinicians read chain-of-thought (CoT) rationales as evidence of medical reasoning, but whether the visible chain plays that role is rarely tested. General-domain CoT-faithfulness probes ignore clinical cost, and medical LLM evaluations treat the chain as a black box. We close this gap with a medical perturbation audit: a 30-operator battery edits both the chain and the question with clinically motivated operators (severity reversal, negation flip, demographic swap, evidence ablation), paired with a chain-update times answer-flip joint analysis that classifies each model by its failure mode. A

Read original source ↗ Open in workspace

recordType

paper

region

Global