MARS: Multi-Agent Adaptive Reasoning with Socratic Guidance for Automated Prompt Optimization

Jian Zhang, Zhangqi Wang, Haiping Zhu +6

Large language models (LLMs) typically operate in a question-answering paradigm, where the quality of the input prompt critically affects the response. Automated Prompt Optimizatio…