Prompt Stability Matters: Evaluating and Optimizing Auto-Generated Prompt in General-Purpose Systems

2025-05-19

AI Summary

Existing prompt generation methods primarily rely on task accuracy or similar outcome-based metrics, neglecting the intrinsic stability—i.e., response consistency across repeated executions—of prompts, leading to poor interpretability and robustness. Method: This paper introduces *prompt semantic stability* as a necessary condition for LLM prompt reliability and proposes the first stability-aware iterative prompt generation framework. It features (i) an automated stability evaluation mechanism grounded in multi-execution consistency and semantic similarity; (ii) a feedback-driven optimization pipeline; and (iii) a fine-tuned LLaMA-based dedicated stability evaluator. Contribution/Results: Evaluated on both general-purpose and domain-specific tasks, the framework significantly improves both task accuracy and output consistency. Empirical results demonstrate that incorporating stability feedback systematically enhances prompt quality and end-to-end execution reliability.

Technology Category

• Natural Language Processing: Prompt Engineering / Prompting
• Machine Learning: Large Multimodal Models (LMMs)
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Automatic prompt generation plays a crucial role in enabling general-purpose multi-agent systems to perform diverse tasks autonomously. Existing methods typically evaluate prompts based on their immediate task performance, overlooking the intrinsic qualities that determine their reliability. This outcome-centric view not only limits interpretability but also fails to account for the inherent stochasticity of large language models (LLMs). In this work, we bring attention to prompt stability-the consistency of model responses across repeated executions-as a key factor for building robust and effective prompt generation systems. To quantify this, we propose semantic stability as a criterion for assessing the response consistency of prompts, and fine-tune a LLaMA-based evaluator to measure it automatically across tasks. These components have enabled us to develop the first stability-aware general-purpose prompt generation system that leverages stability feedback to iteratively enhance both prompt quality and system-level performance. Furthermore, we establish a logical chain between prompt stability and task success by analyzing the structural dependencies within our system, proving stability as a necessary condition for effective system-level execution. Empirical results across general and domain-specific tasks demonstrate that our stability-aware framework improves both accuracy and output consistency. By shifting the focus from one-off results to persistent reliability, our work offers a new perspective on prompt design and contributes practical tools for building more trustworthy general-purpose systems.

Problem

• Evaluating prompt stability in auto-generated prompts for reliability
• Proposing semantic stability to measure response consistency across tasks
• Developing stability-aware prompt generation to improve accuracy and consistency

Innovation

• Proposes semantic stability for prompt evaluation
• Fine-tunes LLaMA-based evaluator for automatic measurement
• Develops stability-aware prompt generation system