Notes — Multi-modal LLMs for Medical Imaging Published: January 05, 2025Key IdeasVision-language alignment needs clinical context supervision.Reporting-style training targets may improve factual consistency.