Reproduce “Decision language models as affordable and rapid alternatives to LLM judges of clinical generative AI” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Rotem Gershon, Seffi Cohen, Dean Ariel, Havatzelet Bilavsky-Yarden, Amiel Sberro, Shai Yitzhaki, Osnat Tausky, Naama Golan and 6 more
The paper evaluates Jev, a decision language model that returns structured scores or yes/no judgments in parallel for recurring clinical-AI evaluation. On 118 answers from a pediatric decision-support system, its expected ordinal score achieved Gwet’s AC2 of 0.53 against five physicians, matching or exceeding seven of 12 LLM judges. Five scored higher; the study measures agreement, cost, and time, not clinical benefit.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.