arxivcs.CLcs.AI2026-07-21
MIRA-Ev:A Benchmark for Granular Evidence Detection and Relational Reasoning in Clinical Exams
Iker De la Iglesia, Johanna Ramirez-Romero, Jose Maria Villa-Gonzalez, Irune Urroz García, Ander Barrena, Aitziber Atutxa
Clinical NLP evaluation remains dominated by multiple-choice question answering (MCQA), which scores only final-answer accuracy and cannot detect when a model reaches the correct diagnosis while grounding it in irrelevant, absent, or contradictory evidence. We introduce MIRA-Ev,…