Application of unified health large language model evaluation framework to In-Basket message replies: bridging qualitative and quantitative assessments

A unified evaluation framework is proposed that bridges the gap between qualitative human assessments and automated quantitative metrics, enhancing the reliability and scalability of LLM evaluations in healthcare.