<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="choi-kim-2026-interactive">
    <titleInfo>
        <title>Interactive and Scalable Evaluation of LLM Math Tutoring via Simulated Teacher-Student Dialogues</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Gihyeon</namePart>
        <namePart type="family">Choi</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Harksoo</namePart>
        <namePart type="family">Kim</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-11</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Yue</namePart>
            <namePart type="family">Zhang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ruifeng</namePart>
            <namePart type="family">Xu</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ayu</namePart>
            <namePart type="family">Purwarianti</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Tong</namePart>
            <namePart type="family">Rong</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jun</namePart>
            <namePart type="family">Suzuki</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Hengqin, China</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-444-6</identifier>
    </relatedItem>
    <abstract>Evaluating LLMs as math tutors is challenging because many automated protocols either score a single response against a reference or place the model in a fixed dialogue context, without measuring its ability to build and steer tutoring interactions. We present a fully automated, simulation-based framework that generates controlled multi-turn teacher–student dialogues with turn-level answer verification under a fixed interaction budget. This budget specifies the maximum number of student–teacher rounds allowed per problem (K=5 in our main experiments). To prevent score inflation from answer-revealing feedback, we introduce an overinformative eligibility filter: a hint-only student test applied to the teacher utterance immediately preceding the student’s first correct attempt. Only eligible corrections receive credit in our effectiveness and efficiency metrics, both of which are normalized by the number of initially incorrect sessions. We also propose a diagnostic metric suite separating tutoring effectiveness/efficiency (ECR/FEI), answer reachability (ARG), and teacher judgment (correctness/incorrectness F1). Experiments on GSM8K and MATH with diverse teacher and student models show that problem-solving accuracy is an informative but incomplete proxy for interactive tutoring ability, and that filtering overinformative feedback can substantially alter apparent performance. Our framework enables scalable, reproducible, and fine-grained evaluation of LLM math tutoring beyond outcome-only scoring.</abstract>
    <identifier type="citekey">choi-kim-2026-interactive</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-aacl.166/</url>
    </location>
    <part>
        <date>2026-11</date>
        <extent unit="page">
            <start>3300</start>
            <end>3313</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Interactive and Scalable Evaluation of LLM Math Tutoring via Simulated Teacher-Student Dialogues
%A Choi, Gihyeon
%A Kim, Harksoo
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-444-6
%F choi-kim-2026-interactive
%X Evaluating LLMs as math tutors is challenging because many automated protocols either score a single response against a reference or place the model in a fixed dialogue context, without measuring its ability to build and steer tutoring interactions. We present a fully automated, simulation-based framework that generates controlled multi-turn teacher–student dialogues with turn-level answer verification under a fixed interaction budget. This budget specifies the maximum number of student–teacher rounds allowed per problem (K=5 in our main experiments). To prevent score inflation from answer-revealing feedback, we introduce an overinformative eligibility filter: a hint-only student test applied to the teacher utterance immediately preceding the student’s first correct attempt. Only eligible corrections receive credit in our effectiveness and efficiency metrics, both of which are normalized by the number of initially incorrect sessions. We also propose a diagnostic metric suite separating tutoring effectiveness/efficiency (ECR/FEI), answer reachability (ARG), and teacher judgment (correctness/incorrectness F1). Experiments on GSM8K and MATH with diverse teacher and student models show that problem-solving accuracy is an informative but incomplete proxy for interactive tutoring ability, and that filtering overinformative feedback can substantially alter apparent performance. Our framework enables scalable, reproducible, and fine-grained evaluation of LLM math tutoring beyond outcome-only scoring.
%U https://aclanthology.org/2026.findings-aacl.166/
%P 3300-3313

Markdown (Informal)

[Interactive and Scalable Evaluation of LLM Math Tutoring via Simulated Teacher-Student Dialogues](https://aclanthology.org/2026.findings-aacl.166/) (Choi & Kim, Findings 2026)

• Interactive and Scalable Evaluation of LLM Math Tutoring via Simulated Teacher-Student Dialogues (Choi & Kim, Findings 2026)

ACL

• Gihyeon Choi and Harksoo Kim. 2026. Interactive and Scalable Evaluation of LLM Math Tutoring via Simulated Teacher-Student Dialogues. In Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026, pages 3300–3313, Hengqin, China. Association for Computational Linguistics.