<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ashizawa-etal-2025-bandit">
    <titleInfo>
        <title>Bandit-Based Prompt Design Strategy Selection Improves Prompt Optimizers</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Rin</namePart>
        <namePart type="family">Ashizawa</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yoichi</namePart>
        <namePart type="family">Hirose</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Nozomu</namePart>
        <namePart type="family">Yoshinari</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Kento</namePart>
        <namePart type="family">Uchida</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Shinichi</namePart>
        <namePart type="family">Shirakawa</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2025-07</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: ACL 2025</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Wanxiang</namePart>
            <namePart type="family">Che</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Joyce</namePart>
            <namePart type="family">Nabende</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ekaterina</namePart>
            <namePart type="family">Shutova</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Mohammad</namePart>
            <namePart type="given">Taher</namePart>
            <namePart type="family">Pilehvar</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Vienna, Austria</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-256-5</identifier>
    </relatedItem>
    <abstract>Prompt optimization aims to search for effective prompts that enhance the performance of large language models (LLMs). Although existing prompt optimization methods have discovered effective prompts, they often differ from sophisticated prompts carefully designed by human experts. Prompt design strategies, representing best practices for improving prompt performance, can be key to improving prompt optimization. Recently, a method termed the Autonomous Prompt Engineering Toolbox (APET) has incorporated various prompt design strategies into the prompt optimization process. In APET, the LLM is needed to implicitly select and apply the appropriate strategies because prompt design strategies can have negative effects. This implicit selection may be suboptimal due to the limited optimization capabilities of LLMs. This paper introduces Optimizing Prompts with sTrategy Selection (OPTS), which implements explicit selection mechanisms for prompt design. We propose three mechanisms, including a Thompson sampling-based approach, and integrate them into EvoPrompt, a well-known prompt optimizer. Experiments optimizing prompts for two LLMs, Llama-3-8B-Instruct and GPT-4o mini, were conducted using BIG-Bench Hard. Our results show that the selection of prompt design strategies improves the performance of EvoPrompt, and the Thompson sampling-based mechanism achieves the best overall results. Our experimental code is provided at https://github.com/shiralab/OPTS.</abstract>
    <identifier type="citekey">ashizawa-etal-2025-bandit</identifier>
    <identifier type="doi">10.18653/v1/2025.findings-acl.1070</identifier>
    <location>
        <url>https://aclanthology.org/2025.findings-acl.1070/</url>
    </location>
    <part>
        <date>2025-07</date>
        <extent unit="page">
            <start>20799</start>
            <end>20817</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Bandit-Based Prompt Design Strategy Selection Improves Prompt Optimizers
%A Ashizawa, Rin
%A Hirose, Yoichi
%A Yoshinari, Nozomu
%A Uchida, Kento
%A Shirakawa, Shinichi
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Findings of the Association for Computational Linguistics: ACL 2025
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-256-5
%F ashizawa-etal-2025-bandit
%X Prompt optimization aims to search for effective prompts that enhance the performance of large language models (LLMs). Although existing prompt optimization methods have discovered effective prompts, they often differ from sophisticated prompts carefully designed by human experts. Prompt design strategies, representing best practices for improving prompt performance, can be key to improving prompt optimization. Recently, a method termed the Autonomous Prompt Engineering Toolbox (APET) has incorporated various prompt design strategies into the prompt optimization process. In APET, the LLM is needed to implicitly select and apply the appropriate strategies because prompt design strategies can have negative effects. This implicit selection may be suboptimal due to the limited optimization capabilities of LLMs. This paper introduces Optimizing Prompts with sTrategy Selection (OPTS), which implements explicit selection mechanisms for prompt design. We propose three mechanisms, including a Thompson sampling-based approach, and integrate them into EvoPrompt, a well-known prompt optimizer. Experiments optimizing prompts for two LLMs, Llama-3-8B-Instruct and GPT-4o mini, were conducted using BIG-Bench Hard. Our results show that the selection of prompt design strategies improves the performance of EvoPrompt, and the Thompson sampling-based mechanism achieves the best overall results. Our experimental code is provided at https://github.com/shiralab/OPTS.
%R 10.18653/v1/2025.findings-acl.1070
%U https://aclanthology.org/2025.findings-acl.1070/
%U https://doi.org/10.18653/v1/2025.findings-acl.1070
%P 20799-20817

Markdown (Informal)

[Bandit-Based Prompt Design Strategy Selection Improves Prompt Optimizers](https://aclanthology.org/2025.findings-acl.1070/) (Ashizawa et al., Findings 2025)

• Bandit-Based Prompt Design Strategy Selection Improves Prompt Optimizers (Ashizawa et al., Findings 2025)

ACL

• Rin Ashizawa, Yoichi Hirose, Nozomu Yoshinari, Kento Uchida, and Shinichi Shirakawa. 2025. Bandit-Based Prompt Design Strategy Selection Improves Prompt Optimizers. In Findings of the Association for Computational Linguistics: ACL 2025, pages 20799–20817, Vienna, Austria. Association for Computational Linguistics.