No Loss, No Gain: Gated Refinement and Adaptive Compression for Prompt Optimization

2025-09-27

AI Summary

Existing automated prompt optimization methods suffer from premature convergence to local optima, poor stability, and low efficiency. To address these limitations, we propose GRACE, a novel framework featuring gated refinement—comprising a feedback-regulation gate and an update-rejection gate—coupled with adaptive lossy compression. When prompt optimization stagnates, GRACE proactively reconstructs the optimization trajectory, uniquely transforming controlled information loss into optimization gain. Furthermore, GRACE integrates core-concept distillation with structural reorganization to enable dynamic, robust prompt evolution. Evaluated across 11 diverse tasks, GRACE achieves an average relative improvement of 4.7% over prior state-of-the-art methods. Notably, it attains superior performance using only 25% of the prompt-generation budget required by competing approaches, thereby substantially reducing computational overhead.

Technology Category

• Search and Optimization: Learning to Search
• Natural Language Processing: Prompt Engineering / Prompting
• Machine Learning: Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Prompt engineering is crucial for leveraging the full potential of large language models (LLMs). While automatic prompt optimization offers a scalable alternative to costly manual design, generating effective prompts remains challenging. Existing methods often struggle to stably generate improved prompts, leading to low efficiency, and overlook that prompt optimization easily gets trapped in local optima. Addressing this, we propose GRACE, a framework that integrates two synergistic strategies: Gated Refinement and Adaptive Compression, achieving Efficient prompt optimization. The gated refinement strategy introduces a feedback regulation gate and an update rejection gate, which refine update signals to produce stable and effective prompt improvements. When optimization stagnates, the adaptive compression strategy distills the prompt's core concepts, restructuring the optimization trace and opening new paths. By strategically introducing information loss through refinement and compression, GRACE delivers substantial gains in performance and efficiency. In extensive experiments on 11 tasks across three practical domains, including BIG-Bench Hard (BBH), domain-specific, and general NLP tasks, GRACE achieves significant average relative performance improvements of 4.7%, 4.4% and 2.7% over state-of-the-art methods, respectively. Further analysis shows that GRACE achieves these gains using only 25% of the prompt generation budget required by prior methods, highlighting its high optimization efficiency and low computational overhead. Our code is available at https://github.com/Eric8932/GRACE.

Problem

• Optimizing prompts for large language models efficiently and stably
• Overcoming local optima in automatic prompt optimization processes
• Reducing computational overhead while improving prompt performance

Innovation

• Gated refinement strategy regulates prompt update signals
• Adaptive compression distills core concepts when stagnating
• Strategic information loss enables efficient prompt optimization