SteeraMed-Toxicity: a benchmark for language models ranking rare, unexpected adverse effects — the semaglutide-NAION case, long-tail label recall, and dual-gold-standard gene nomination. Paper: preprint forthcoming.
benchmark drug-repurposing clinical-trials drug-safety long-tail toxicity pharmacovigilance large-language-models llm-evaluation adverse-effects glp-1 semaglutide
-
Updated
Oct 5, 2026