BioConflict: A Benchmark for Evaluating Large Language Models in Biomedical Contradiction Detection and Consensus Synthesis
Ashwin Kirubakaran, Henry Gagnier
BioNLP @ ACL 2026 · pp. 552–558 · 2026.bionlp-1.44
250 expert-annotated paper pairs for detecting conflicts between biomedical papers, extracting the variables behind them, and synthesizing a consensus.