SEE: Strategic Exploration and Exploitation for Cohesive In-Context Prompt Optimization

TL;DR

This paper introduces SEE, a scalable framework for optimizing prompts in Large Language Models by balancing exploration and exploitation, leading to more cohesive prompts and improved task performance.

Contribution

The paper presents a novel metaheuristic-based optimization framework that refines both instructions and examples for prompt design, addressing previous incohesiveness and efficiency issues.

Findings

SEE outperforms baseline methods by 13.94 in average performance.

SEE reduces computational costs by 58.67.

Evaluation across 35 benchmark tasks demonstrates significant improvements.

Abstract

Designing optimal prompts for Large Language Models (LLMs) is a complicated and resource-intensive task, often requiring substantial human expertise and effort. Existing approaches typically separate the optimization of prompt instructions and in-context learning examples, leading to incohesive prompts that are defined and represented by suboptimal task performance. To overcome these challenges, we propose a novel Cohesive In-Context Prompt Optimization framework that refines both prompt instructions and examples. However, formulating such an optimization in the discrete and high-dimensional space of natural language poses significant challenges in both convergence and computational efficiency. To address these issues, we introduce SEE, a scalable and efficient prompt optimization framework that adopts metaheuristic optimization principles and strategically balances exploration and exploitation to enhance optimization performance and achieve efficient convergence. SEE features a quad-phased design that alternates between global traversal (exploration) and local optimization (exploitation) and adaptively chooses LLM operators during the optimization process. We have conducted a comprehensive evaluation across 35 benchmark tasks, and SEE significantly outperforms state-of-the-art baseline methods by a large margin, achieving an average performance gain of 13.94 while reducing computational costs by 58.67.