Skip to main content
Loading ResearchHub…
A comparative evaluation of a domain-specific large language model and GPT-4o for radiology report summarisation | ResearchHub