<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="knezevic-etal-2026-evaluating">
    <titleInfo>
        <title>Evaluating Inquiry Learning in LLM-Based Agents: Hypothesizing, Reasoning, and Exploration</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Marta</namePart>
        <namePart type="family">Knežević</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Seyed</namePart>
        <namePart type="given">Parsa</namePart>
        <namePart type="family">Neshaei</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jibril</namePart>
        <namePart type="family">Frej</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Vinitra</namePart>
        <namePart type="family">Swamy</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Tanja</namePart>
        <namePart type="family">Käser</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Hung-Yi</namePart>
            <namePart type="family">Lee</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Sunipa</namePart>
            <namePart type="family">Dev</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jessy</namePart>
            <namePart type="family">Li</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Desmond</namePart>
            <namePart type="family">Elliott</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-462-0</identifier>
    </relatedItem>
    <abstract>Scientific inquiry involves iterative cycles of exploration, hypothesis generation, and reasoning over observations to uncover relationships in complex systems. Large language models (LLMs) have demonstrated strong capabilities in several components of this process, including inductive reasoning, hypothesis generation, and data analysis. However, prior work typically evaluates these abilities in isolation, leaving open how models integrate them across a full inquiry cycle. In this work, we introduce a framework for evaluating LLMs’ inquiry capabilities in a unified, interactive setting. We extend the Countdown task to a mathematical environment where an agent probes the system by manipulating variables, collecting observations, and inferring relationships between them. This setup enables analysis of exploration strategies, causal discovery, and hypothesis refinement during inquiry. Across seven LLMs and multiple prompting strategies, our experiments reveal limited inherent inquiry capabilities and low success rates in discovering the underlying expressions. However, structured guidance inspired by educational scaffolding significantly improves performance. Our findings highlight current limitations of LLMs in strategic exploration and iterative hypothesis refinement, and show how simple guidance strategies developed for human learners in inquiry learning can support more effective inquiry behavior.</abstract>
    <identifier type="citekey">knezevic-etal-2026-evaluating</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.1861/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>39716</start>
            <end>39737</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Evaluating Inquiry Learning in LLM-Based Agents: Hypothesizing, Reasoning, and Exploration
%A Knežević, Marta
%A Neshaei, Seyed Parsa
%A Frej, Jibril
%A Swamy, Vinitra
%A Käser, Tanja
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F knezevic-etal-2026-evaluating
%X Scientific inquiry involves iterative cycles of exploration, hypothesis generation, and reasoning over observations to uncover relationships in complex systems. Large language models (LLMs) have demonstrated strong capabilities in several components of this process, including inductive reasoning, hypothesis generation, and data analysis. However, prior work typically evaluates these abilities in isolation, leaving open how models integrate them across a full inquiry cycle. In this work, we introduce a framework for evaluating LLMs’ inquiry capabilities in a unified, interactive setting. We extend the Countdown task to a mathematical environment where an agent probes the system by manipulating variables, collecting observations, and inferring relationships between them. This setup enables analysis of exploration strategies, causal discovery, and hypothesis refinement during inquiry. Across seven LLMs and multiple prompting strategies, our experiments reveal limited inherent inquiry capabilities and low success rates in discovering the underlying expressions. However, structured guidance inspired by educational scaffolding significantly improves performance. Our findings highlight current limitations of LLMs in strategic exploration and iterative hypothesis refinement, and show how simple guidance strategies developed for human learners in inquiry learning can support more effective inquiry behavior.
%U https://aclanthology.org/2026.findings-emnlp.1861/
%P 39716-39737

Markdown (Informal)

[Evaluating Inquiry Learning in LLM-Based Agents: Hypothesizing, Reasoning, and Exploration](https://aclanthology.org/2026.findings-emnlp.1861/) (Knežević et al., Findings 2026)

• Evaluating Inquiry Learning in LLM-Based Agents: Hypothesizing, Reasoning, and Exploration (Knežević et al., Findings 2026)

ACL

• Marta Knežević, Seyed Parsa Neshaei, Jibril Frej, Vinitra Swamy, and Tanja Käser. 2026. Evaluating Inquiry Learning in LLM-Based Agents: Hypothesizing, Reasoning, and Exploration. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 39716–39737, Budapest, Hungary. Association for Computational Linguistics.