Enhancing Large Language Models for Improved Accuracy and Safety in Medical Question Answering: Comparative Study
The Med-RISE framework significantly improves the accuracy and reduces the hallucinations of LLMs in medical question answering across benchmark datasets. By providing local and real-time information retrieval and fact and safety filtering, Med-RISE enhances the reliability and interpretability of LLMs in the medical domain, offering a promising tool for clinical practice and decision support.
Authors
- Caineng Pan (ORCID: https://orcid.org/0000-0001-7988-5363)
- Yingfeng Zheng (ORCID: https://orcid.org/0000-0002-9952-6445)
- Mengxiang Guo (ORCID: https://orcid.org/0000-0002-5974-3533)
- Jingni Li (ORCID: https://orcid.org/0009-0006-4716-1532)
- Z Y Liu (ORCID: https://orcid.org/0000-0001-9209-4179)
- Wen Shi (ORCID: https://orcid.org/0009-0006-5084-5265)
- Jinguo Ye
- Dingqiao Wang (ORCID: https://orcid.org/0009-0007-2648-6677)
- Dongliang Wang (ORCID: https://orcid.org/0000-0003-0605-5726)
- Wei Du (ORCID: https://orcid.org/0009-0006-5558-1625)
- Qikai Zhang (ORCID: https://orcid.org/0009-0002-3183-3200)
- J. K. Liang (ORCID: https://orcid.org/0009-0002-6290-5748)
- Fei Li (ORCID: https://orcid.org/0000-0001-6211-6306)
- Qiuling Hu (ORCID: https://orcid.org/0009-0004-5627-776X)
Institutions
- Sun Yat-sen University (CN)
- Guangzhou Women and Children Medical Center (CN)
- Eighth Affiliated Hospital of Sun Yat-sen University
- Guangzhou Medical University (CN)
Publication Details
- Journal
- JMIR Medical Education
- Published
- 2025-09-30
- DOI
- https://doi.org/10.2196/70190
- Citations
- 2
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 1.31