Rethinking Prompt Optimizers: From Prompt Merits to Optimization

TL;DR

This paper proposes MePO, an interpretable, merit-guided prompt optimizer trained on a lightweight LLM dataset, which improves prompt quality and response accuracy across various models without online optimization.

Contribution

It introduces a novel, model-agnostic prompt optimization method based on explicit quality merits, enhancing interpretability and generalization for diverse LLMs.

Findings

MePO outperforms existing prompt optimization methods across multiple tasks.

It effectively generalizes to both large-scale and lightweight models.

The approach reduces privacy concerns by avoiding online optimization.

Abstract

Prompt optimization (PO) provides a practical way to improve response quality when users lack the time or expertise to manually craft effective prompts. Existing methods typically rely on LLMs' self-generation ability to optimize prompts. However, due to limited downward compatibility, the instruction-heavy prompts generated by advanced LLMs can overwhelm lightweight inference models and degrade response quality, while also lacking interpretability due to implicit optimization. In this work, we rethink prompt optimization through the lens of explicit and interpretable design. We first identify a set of model-agnostic prompt quality merits and empirically validate their effectiveness in enhancing prompt and response quality. We then introduce MePO, a merit-guided, locally deployable prompt optimizer trained on our merit-guided prompt preference dataset generated by a lightweight LLM. MePO avoids online optimization, reduces privacy concerns, and, by learning clear, interpretable merits, generalizes effectively to both large-scale and lightweight inference models. Experiments demonstrate that MePO achieves better results across diverse tasks and model types, offering a scalable and robust solution for real-world deployment. The code, model and dataset can be found in https://github.com/MidiyaZhu/MePO