[pkroll]

https://arxiv.org/abs/2412.10849

"Our study suggests that LLMs have achieved superhuman performance on general medical diagnostic and management reasoning"

[squigz]

This isn't really applying LLMs to research in novel ways.