Automatic Prompt Optimization via Heuristic Search: A Survey

Wendi Cui, Jiaxin Zhang, Zhuohang Li, Hao Sun, Damien Lopez, Kamalika Das, Bradley Malin, Sricharan Kumar · 2025

Read Automatic Prompt Optimization via Heuristic Search: A Survey by Wendi Cui, Jiaxin Zhang, Zhuohang Li, Hao Sun, Damien Lopez, Kamalika Das, Bradley Malin, Sricharan Kumar (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-18746.txt.

Description

Wendi Cui and co-authors survey automatic prompt optimization in a paper from 2025. The head frames it as a survey of heuristic search for instruction-focused prompts. Prompts decide how well language models perform, yet hand tuning leans on gut feel and never improves itself. Search algorithms can refine prompts on their own.

The survey sorts methods by five questions: where the search acts, what it changes, what score guides it, what operators propose new prompts, and which search loop runs it. Special datasets and tools that speed the work get their own tour. The taxonomy gives builders a menu of proven designs. Automatic tuning may soon be as standard as model fine-tuning.

Catalog record

CREATORWendi Cui, Jiaxin Zhang, Zhuohang Li, Hao Sun, Damien Lopez, Kamalika Das, Bradley Malin, Sricharan Kumar
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE32 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-18746

Browse the full library · About the Tower of Babel