vix.ing · top · new · best · stats

Can large language models reason about medical questions?

2022/07/17 by Valentin Liévin, Christoffer Egeberg Hother, Christoffer Hother +6 · 2 voices · 86 citations
Computer Science · Medicine · Psychology · #Annotation #Artificial Intelligence in Healthcare and Education #Artificial intelligence #Closing (real estate) #Cognitive psychology #Computer science #Expert system #Machine Learning in Healthcare #Natural language processing #Programming language #Psychology #Recall #Set (abstract data type) #Shot (pellet) #Subject-matter expert #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2207.08143

published in arXiv (Cornell University) (Cornell University)

openalex publication_date 2022/07/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Although large language models (LLMs) often produce impressive outputs, it remains unclear how they perform in real-world scenarios requiring strong reasoning skills and expert domain knowledge. We set out to investigate whether close- and open-source models (GPT-3.5, LLama-2, etc.) can be applied to answer and reason about difficult real-world-based questions. We focus on three popular medical benchmarks (MedQA-USMLE, MedMCQA, and PubMedQA) and multiple prompting scenarios: Chain-of-Thought (CoT, think step-by-step), few-shot and retrieval augmentation. Based on an expert annotation of the generated CoTs, we found that InstructGPT can often read, reason and recall expert knowledge. Last, by leveraging advances in prompt engineering (few-shot and ensemble methods), we demonstrated that GPT-3.5 not only yields calibrated predictive distributions, but also reaches the passing score on three datasets: MedQA-USMLE 60.2%, MedMCQA 62.7% and PubMedQA 78.2%. Open-source models are closing the gap: Llama-2 70B also passed the MedQA-USMLE with 62.5% accuracy.

Cited by

Discussions

Related