Healthcare-Specific Large Language Models: Capabilities & Benchmarks
Wednesday, June 21, 2023
Red Rock Ballroom I
Large language models like GPT-4 and their open-source counterparts provide a leap in capabilities on understanding medical language and context – from passing the US medical licensing exam to summarizing clinical notes. Recently, a wave of health-specific large language models is showing that tuning models specifically on medical data and tasks can result in even higher accuracy on common use cases such as question answering, information extraction, and summarization. Some of these models also aim to address the privacy, hallucination, and fairness issues that current language models exhibit. This session reviews the current state of the art and provides recommendations on what to consider when deploying these technologies in practice.