GRAD-SUM: Leveraging Gradient Summarization for Optimal Prompt Engineering

TL;DR

GRAD-SUM is a scalable, gradient-based method for automatic prompt engineering that effectively optimizes prompts for large language models, outperforming existing approaches across multiple benchmarks.

Contribution

The paper introduces GRAD-SUM, a novel gradient summarization technique for automatic prompt optimization that is flexible, task-agnostic, and more effective than prior methods.

Findings

Outperforms existing prompt engineering methods on various benchmarks.

Incorporates user-defined task descriptions and evaluation criteria.

Demonstrates versatility and effectiveness in automatic prompt optimization.

Abstract

Prompt engineering for large language models (LLMs) is often a manual time-intensive process that involves generating, evaluating, and refining prompts iteratively to ensure high-quality outputs. While there has been work on automating prompt engineering, the solutions generally are either tuned to specific tasks with given answers or are quite costly. We introduce GRAD-SUM, a scalable and flexible method for automatic prompt engineering that builds on gradient-based optimization techniques. Our approach incorporates user-defined task descriptions and evaluation criteria, and features a novel gradient summarization module to generalize feedback effectively. Our results demonstrate that GRAD-SUM consistently outperforms existing methods across various benchmarks, highlighting its versatility and effectiveness in automatic prompt optimization.