O37: Demystifying base large language model: Reproducibility and accuracy in ACMG/AMP variant classification
Abstract
Large language models (LLMs) are increasingly used by both consumers and clinicians to interpret medical and genetic test results via AI-powered chatbots and inference engines. Recent work underscores concern about their reproducibility, safety, and parameter sensitivity in these settings. Although online chatbots are not direct model calls but wrap foundation models via complex structure with multiple API calls and other machine learning models, the intrinsic nature of generative models exhibits random variation.
Fields of Study & Subjects
Artificial Intelligence in Healthcare and Education
Genomics and Rare Diseases
Machine Learning in Healthcare
Publication details
- Journal / source
- Genetics in Medicine Open
- Volume
- 4
- Pages
- 104325–104325
- Publication date
- 1 January 2026
- Language
- EN