Tell Me You're Biased Without Telling Me You're Biased -- Toward Revealing Implicit Biases in Medical LLMs
This addresses the crucial need to detect implicit biases in medical LLMs before clinical deployment, though it appears incremental as it builds on existing bias detection methods with specific enhancements.
The researchers tackled the problem of identifying biased patterns in medical large language models (LLMs) by developing a novel framework that combines knowledge graphs with auxiliary LLMs and adversarial perturbations, showing it has noticeably greater ability and scalability to reveal complex biases compared to baselines across three datasets, six LLMs, and five bias types.
Large language models (LLMs) that are used in medical applications are known to show biased and unfair patterns. Prior to adopting these in clinical decision-making applications, it is crucial to identify these bias patterns to enable effective mitigation of their impact. In this study, we present a novel framework combining knowledge graphs (KGs) with auxiliary LLMs to systematically reveal complex bias patterns in medical LLMs. Specifically, the proposed approach integrates adversarial perturbation techniques to identify subtle bias patterns. The approach adopts a customized multi-hop characterization of KGs to enhance the systematic evaluation of arbitrary LLMs. Through a series of comprehensive experiments (on three datasets, six LLMs, and five bias types), we show that our proposed framework has noticeably greater ability and scalability to reveal complex biased patterns of LLMs compared to other baselines.