Localized Zeroth-Order Prompt Optimization
This work addresses the challenge of optimizing prompts for black-box LLMs, offering a more efficient approach for users relying on prompt-based methods, though it is incremental in improving existing optimization techniques.
The paper tackles the problem of inefficient global optimization in prompt-based methods for large language models by showing that local optima are often sufficient and more efficient to find. The proposed localized zeroth-order prompt optimization (ZOPO) algorithm outperforms existing baselines in optimization performance and query efficiency, as demonstrated through extensive experiments.
The efficacy of large language models (LLMs) in understanding and generating natural language has aroused a wide interest in developing prompt-based methods to harness the power of black-box LLMs. Existing methodologies usually prioritize a global optimization for finding the global optimum, which however will perform poorly in certain tasks. This thus motivates us to re-think the necessity of finding a global optimum in prompt optimization. To answer this, we conduct a thorough empirical study on prompt optimization and draw two major insights. Contrasting with the rarity of global optimum, local optima are usually prevalent and well-performed, which can be more worthwhile for efficient prompt optimization (Insight I). The choice of the input domain, covering both the generation and the representation of prompts, affects the identification of well-performing local optima (Insight II). Inspired by these insights, we propose a novel algorithm, namely localized zeroth-order prompt optimization (ZOPO), which incorporates a Neural Tangent Kernel-based derived Gaussian process into standard zeroth-order optimization for an efficient search of well-performing local optima in prompt optimization. Remarkably, ZOPO outperforms existing baselines in terms of both the optimization performance and the query efficiency, which we demonstrate through extensive experiments.