<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lee-etal-2026-foundation">
    <titleInfo>
        <title>Foundation First: Stabilizing SLM Agents via Pre-execution Logical Scaffolding</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Juntae</namePart>
        <namePart type="family">Lee</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jihwan</namePart>
        <namePart type="family">Bang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Seunghan</namePart>
        <namePart type="family">Yang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Sungha</namePart>
        <namePart type="family">Choi</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>Proceedings of the Second Workshop for REsearch on Agent Language Models (REALM 2026)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Ehsan</namePart>
            <namePart type="family">Kamalloo</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Nicolas</namePart>
            <namePart type="family">Gontier</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Nikolai</namePart>
            <namePart type="family">Rozanov</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Maryam</namePart>
            <namePart type="family">Hashemzadeh</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Akhil</namePart>
            <namePart type="family">Arora</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">979-8-89176-485-9</identifier>
    </relatedItem>
    <abstract>In on-device agentic systems, Small Language Models (SLMs) are essential for ensuring privacy-preserving personalization and low-latency execution. However, despite their capacity to handle isolated atomic tasks, during complex multi-step orchestration SLMs frequently suffer from stochastic drift—a phenomenon where minor logical deviations cascade into irrecoverable trajectory failures. In this work, we demonstrate the primacy of pre-execution planning, revealing that mitigating initial logical drift is vastly more effective than mid-trajectory interventions. Building on this insight, we propose FORST(FOundation fiRST) framework that implements sequential logical scaffolding in pre-execution planning, where predefined expert anchors are injected as fixed pre-computed KV-cache states. While the scaffolding ensures logical stability, KV-cache reuse efficiently realizes the structured anchor prefix without repeatedly prefilling static anchors. Then, we can maintain a monolithic reasoning chain within predefined logical planning boundaries and effectively eliminate the redundant computational overhead of structured anchors. Extensive evaluations on the MCPEval benchmark demonstrate that establishing this rigid logical foundation via achieves competitive or numerically higher performance than frontier models, while achieving an latency reduction on edge-device platforms.</abstract>
    <identifier type="citekey">lee-etal-2026-foundation</identifier>
    <location>
        <url>https://aclanthology.org/2026.realm-1.11/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>146</start>
            <end>160</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Foundation First: Stabilizing SLM Agents via Pre-execution Logical Scaffolding
%A Lee, Juntae
%A Bang, Jihwan
%A Yang, Seunghan
%A Choi, Sungha
%Y Kamalloo, Ehsan
%Y Gontier, Nicolas
%Y Rozanov, Nikolai
%Y Hashemzadeh, Maryam
%Y Arora, Akhil
%S Proceedings of the Second Workshop for REsearch on Agent Language Models (REALM 2026)
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-485-9
%F lee-etal-2026-foundation
%X In on-device agentic systems, Small Language Models (SLMs) are essential for ensuring privacy-preserving personalization and low-latency execution. However, despite their capacity to handle isolated atomic tasks, during complex multi-step orchestration SLMs frequently suffer from stochastic drift—a phenomenon where minor logical deviations cascade into irrecoverable trajectory failures. In this work, we demonstrate the primacy of pre-execution planning, revealing that mitigating initial logical drift is vastly more effective than mid-trajectory interventions. Building on this insight, we propose FORST(FOundation fiRST) framework that implements sequential logical scaffolding in pre-execution planning, where predefined expert anchors are injected as fixed pre-computed KV-cache states. While the scaffolding ensures logical stability, KV-cache reuse efficiently realizes the structured anchor prefix without repeatedly prefilling static anchors. Then, we can maintain a monolithic reasoning chain within predefined logical planning boundaries and effectively eliminate the redundant computational overhead of structured anchors. Extensive evaluations on the MCPEval benchmark demonstrate that establishing this rigid logical foundation via achieves competitive or numerically higher performance than frontier models, while achieving an latency reduction on edge-device platforms.
%U https://aclanthology.org/2026.realm-1.11/
%P 146-160

Markdown (Informal)

[Foundation First: Stabilizing SLM Agents via Pre-execution Logical Scaffolding](https://aclanthology.org/2026.realm-1.11/) (Lee et al., REALM 2026)

• Foundation First: Stabilizing SLM Agents via Pre-execution Logical Scaffolding (Lee et al., REALM 2026)

ACL

• Juntae Lee, Jihwan Bang, Seunghan Yang, and Sungha Choi. 2026. Foundation First: Stabilizing SLM Agents via Pre-execution Logical Scaffolding. In Proceedings of the Second Workshop for REsearch on Agent Language Models (REALM 2026), pages 146–160, Budapest, Hungary. Association for Computational Linguistics.