1 paper · 1 filter
Avi-ad Avraam Buskila
Incorporating large language models (LLMs) in medical question answering demands more than high average accuracy: a model that returns substantively different answers each time it…