Limitations of Large Language Models in Clinical Problem-Solving Arising from Inflexible Reasoning

Jonathan Kim, Anna Podlasek, Kie Shidara +3

Large Language Models (LLMs) have attained human-level accuracy on medical question-answer (QA) benchmarks. However, their limitations in navigating open-ended clinical scenarios h…