Proprietary large language models outperform open-source counterparts on multi-choice nephrology questions

Newswise — In the field of natural language processing, it appears that some large language models (LLMs) are better than other artificial intelligence systems when applied to nephrology. A recently published study co-led by scientists at David Geffen School of Medicine at UCLA finds that the proprietary LLMs Claude 2 and GPT-4 outperformed open-source models in correctly answering multiple-choice questions in Nephrology. Read the article in NEJM-AI