<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="alazraki-rei-2025-meta">
    <titleInfo>
        <title>Meta-Reasoning Improves Tool Use in Large Language Models</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Lisa</namePart>
        <namePart type="family">Alazraki</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Marek</namePart>
        <namePart type="family">Rei</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2025-04</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: NAACL 2025</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Luis</namePart>
            <namePart type="family">Chiruzzo</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Alan</namePart>
            <namePart type="family">Ritter</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Lu</namePart>
            <namePart type="family">Wang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Albuquerque, New Mexico</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-195-7</identifier>
    </relatedItem>
    <abstract>External tools help large language models succeed at tasks where they would otherwise typically fail. In existing frameworks, choosing tools at test time relies on naive greedy decoding, regardless of whether the model has been fine-tuned on tool-annotated data or prompted with in-context examples. In contrast, we find that gathering and choosing among a suitable set of candidate tools has greater potential to lead to an optimal selection. We present Tool selECTion via meta-reasONing (TECTON), a two-phase system that first *reasons* over a task and outputs candidate tools using a custom fine-tuned language modelling head. Then, with the custom head disabled, it *meta-reasons* (i.e., it reasons over the previous reasoning process) to make a final choice. We show that TECTON results in substantial gains—both in-distribution and out-of-distribution—on a range of math reasoning datasets.</abstract>
    <identifier type="citekey">alazraki-rei-2025-meta</identifier>
    <identifier type="doi">10.18653/v1/2025.findings-naacl.440</identifier>
    <location>
        <url>https://aclanthology.org/2025.findings-naacl.440/</url>
    </location>
    <part>
        <date>2025-04</date>
        <extent unit="page">
            <start>7885</start>
            <end>7897</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Meta-Reasoning Improves Tool Use in Large Language Models
%A Alazraki, Lisa
%A Rei, Marek
%Y Chiruzzo, Luis
%Y Ritter, Alan
%Y Wang, Lu
%S Findings of the Association for Computational Linguistics: NAACL 2025
%D 2025
%8 April
%I Association for Computational Linguistics
%C Albuquerque, New Mexico
%@ 979-8-89176-195-7
%F alazraki-rei-2025-meta
%X External tools help large language models succeed at tasks where they would otherwise typically fail. In existing frameworks, choosing tools at test time relies on naive greedy decoding, regardless of whether the model has been fine-tuned on tool-annotated data or prompted with in-context examples. In contrast, we find that gathering and choosing among a suitable set of candidate tools has greater potential to lead to an optimal selection. We present Tool selECTion via meta-reasONing (TECTON), a two-phase system that first *reasons* over a task and outputs candidate tools using a custom fine-tuned language modelling head. Then, with the custom head disabled, it *meta-reasons* (i.e., it reasons over the previous reasoning process) to make a final choice. We show that TECTON results in substantial gains—both in-distribution and out-of-distribution—on a range of math reasoning datasets.
%R 10.18653/v1/2025.findings-naacl.440
%U https://aclanthology.org/2025.findings-naacl.440/
%U https://doi.org/10.18653/v1/2025.findings-naacl.440
%P 7885-7897

Markdown (Informal)

[Meta-Reasoning Improves Tool Use in Large Language Models](https://aclanthology.org/2025.findings-naacl.440/) (Alazraki & Rei, Findings 2025)

• Meta-Reasoning Improves Tool Use in Large Language Models (Alazraki & Rei, Findings 2025)

ACL

• Lisa Alazraki and Marek Rei. 2025. Meta-Reasoning Improves Tool Use in Large Language Models. In Findings of the Association for Computational Linguistics: NAACL 2025, pages 7885–7897, Albuquerque, New Mexico. Association for Computational Linguistics.