Dynamic few-shot prompting for clinical note section classification using lightweight, open-source large language models
Our results showcase substantial performance improvements imparted by dynamically selecting examples for few-shot LLM prompting, and further improvement by including section context, demonstrating compelling potential for clinical applications.
Authors
- William Hersh (ORCID: https://orcid.org/0000-0002-4114-5148)
- Andrew Wen (ORCID: https://orcid.org/0000-0001-9090-8028)
- Steven Bedrick (ORCID: https://orcid.org/0000-0002-0163-9397)
- Hongfang Liu (ORCID: https://orcid.org/0000-0003-2570-3741)
- Kirk Roberts (ORCID: https://orcid.org/0000-0003-0716-3630)
- Qiuhao Lu (ORCID: https://orcid.org/0000-0003-2368-8410)
- Kurt Miller
Institutions
- Mayo Clinic (US)
- Oregon Health & Science University (US)
- University of Minnesota Rochester (US)
- The University of Texas Health Science Center at Houston (US)
Publication Details
- Journal
- Journal of the American Medical Informatics Association
- Published
- 2025-06-03
- DOI
- https://doi.org/10.1093/jamia/ocaf084
- Citations
- 2
- Primary Topic
- Machine Learning in Healthcare
- Type
- article
- Field-Weighted Citation Impact
- 4.96
Funders
- National Institutes of Health