Guli Zhu

Evaluating and Understanding Model Editing for Medical Vision Language Models

Guli Zhu, Chenwei Wu, Liyue Shen

Model editing promises a fast, targeted way to correct post-deployment mistakes in medical vision-language models (VLMs) without costly retraining. However, existing multimodal model editing benchmarks focus on general-purpose tasks and do not reflect realistic clinical domain re…Read more