Muchen Yang

Dual-Phase Accelerated Prompt Optimization

Gradient-free prompt optimization methods have made significant strides in enhancing the performance of closed-source Large Language Model (LLMs) across a wide range of tasks. However, existing approaches make light of the importance of high-quality prompt initialization and the identification of ef…