What type of study is this?

This is a Quantitative Study study.

October 8, 2025Open Access

MedHELM: Holistic Evaluation of Large Language Models for Medical Tasks

View Full Paper

Ask AI

Bookmark

Authors

Observational analysis evaluates large language models for medical tasks, indicating significant performance variation.

Evaluation of nine large language models revealed performance variation across multiple medical tasks.
On a normalized accuracy scale, models scored highest in clinical note generation and patient communication.
The study developed a clinician-validated taxonomy with 121 tasks to comprehensively assess model performance.
MedHELM's evaluation framework aligns closely with clinician ratings, suggesting it may improve assessments in medical contexts.

Bedi et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark