Performance evaluation of a large language model for medication management tasks.
Model performance for basic medication tasks was consistently poor, highlighting the need for domain-specific training through clinician-annotated datasets and a comprehensive evaluation framework for benchmarking performance.