Found 1 document(s)
Page 1 of 1
-
Multimodal large language models (MLLMs) can answer clinical questions about medical images, but trustworthy use requires visual grounding: models should localize the pathology they report rather than rely on spurious sh...