Search papers, labs, and topics across Lattice.
This paper introduces CPB-Bench, a new benchmark for evaluating LLMs in medical consultation scenarios, focusing on challenging patient behaviors like information contradiction, factual inaccuracy, self-diagnosis, and care resistance. The benchmark, built on existing medical dialogue datasets, contains 692 multi-turn dialogues annotated with these behaviors in both English and Chinese. Evaluations of various LLMs revealed consistent, behavior-specific failure patterns, particularly when handling contradictory or medically implausible patient information, and tested interventions showed inconsistent improvements.
LLMs struggle to handle common, challenging patient behaviors like contradictory statements and inaccurate medical information, revealing critical safety gaps in medical consultation applications.
Large language models (LLMs) are increasingly used for medical consultation and health information support. In this high-stakes setting, safety depends not only on medical knowledge, but also on how models respond when patient inputs are unclear, inconsistent, or misleading. However, most existing medical LLM evaluations assume idealized and well-posed patient questions, which limits their realism. In this paper, we study challenging patient behaviors that commonly arise in real medical consultations and complicate safe clinical reasoning. We define four clinically grounded categories of such behaviors: information contradiction, factual inaccuracy, self-diagnosis, and care resistance. For each behavior, we specify concrete failure criteria that capture unsafe responses. Building on four existing medical dialogue datasets, we introduce CPB-Bench (Challenging Patient Behaviors Benchmark), a bilingual (English and Chinese) benchmark of 692 multi-turn dialogues annotated with these behaviors. We evaluate a range of open- and closed-source LLMs on their responses to challenging patient utterances. While models perform well overall, we identify consistent, behavior-specific failure patterns, with particular difficulty in handling contradictory or medically implausible patient information. We also study four intervention strategies and find that they yield inconsistent improvements and can introduce unnecessary corrections. We release the dataset and code.
