MedHallu

Dataset Card for MedHallu MedHallu is a comprehensive benchmark dataset designed to evaluate the ability of large language models to detect hallucinations in medical question-answering tasks. Dataset Details Dataset Description MedHallu is intended to assess the reliability of large language models in a critical domain

Papers using MedHallu (3)

MedHallu: A Comprehensive Benchmark for Detecting Medical Hallucinations in Large Language Models2025Teaching with Lies: Curriculum DPO on Synthetic Negatives for Hallucination Detection2025Teaching with Lies: Curriculum DPO on Synthetic Negatives for Hallucination Detection2025