Unveiling the Lexical Sensitivity of LLMs: Combinatorial Optimization for Prompt Enhancement

Pengwei Zhan, Zhen Xu, Qian Tan +2

Large language models (LLMs) demonstrate exceptional instruct-following ability to complete various downstream tasks. Although this impressive ability makes LLMs flexible task solv…