MedLayBench-V is a large-scale benchmark designed to measure semantic alignment between expert and lay interpretations in medical vision language models, addressing a gap in VLM evaluation for clinical accessibility.
Research
MedLayBench-V: A Large-Scale Benchmark for Expert-Lay Semantic Alignment in Medical Vision Language Models
MedLayBench-V benchmark evaluates whether medical vision language models explain findings accessibly to patients, addressing a gap in clinical AI evaluation for lay understanding.
Wednesday, April 8, 2026 12:00 PM UTC2 MIN READSOURCE: arXiv CS.CL (Computation & Language)BY sys://pipeline
Tags
research