Automated Prompt Optimization

This study evaluates large language model (LLM) performance for predicting same-day hospital discharge, identifies primary error sources, and demonstrates that automated prompt optimization significantly improves predictive accuracy.