<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yuan-etal-2026-paths">
    <titleInfo>
        <title>Where Paths Split: Localized, Calibrated Control of Moral Reasoning in Large Language Models</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Chenchen</namePart>
        <namePart type="family">Yuan</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Zheyu</namePart>
        <namePart type="family">Zhang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Gjergji</namePart>
        <namePart type="family">Kasneci</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-07</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Maria</namePart>
            <namePart type="family">Liakata</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Viviane</namePart>
            <namePart type="given">P</namePart>
            <namePart type="family">Moreira</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jiajun</namePart>
            <namePart type="family">Zhang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">David</namePart>
            <namePart type="family">Jurgens</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">San Diego, California, United States</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-390-6</identifier>
    </relatedItem>
    <abstract>Large language models often display heterogeneous moral preferences across settings. We study inference-time steering toward a desired ethical framework while preserving general competence. We present Convergent-Divergent Routing, which traces and edits minimal branch points inside transformer blocks where ethical-framework-related pathways first converge and then diverge. Gating non-target branches at these loci blocks the downstream propagation while leaving upstream computations intact. We find that this intervention alone increases targeted ethical-framework reasoning. To achieve fine-grained control, we adapt Common Spatial Patterns to the residual stream and extract, for each branch-point layer, a pair of directions that discriminate between utilitarian and deontological frameworks. We then introduce Dual Logit Calibration, a closed-form, minimum-\ell₂-norm update that moves the residual within this two-dimensional subspace so the resulting directional projections align with user-specified preference weights. Experiments on real-life moral dilemmas show that our method reliably achieves preference calibration and largely preserves general capabilities, outperforming recent baselines while providing an interpretable mechanism.</abstract>
    <identifier type="citekey">yuan-etal-2026-paths</identifier>
    <location>
        <url>https://aclanthology.org/2026.acl-long.1933/</url>
    </location>
    <part>
        <date>2026-07</date>
        <extent unit="page">
            <start>41698</start>
            <end>41721</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Where Paths Split: Localized, Calibrated Control of Moral Reasoning in Large Language Models
%A Yuan, Chenchen
%A Zhang, Zheyu
%A Kasneci, Gjergji
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-390-6
%F yuan-etal-2026-paths
%X Large language models often display heterogeneous moral preferences across settings. We study inference-time steering toward a desired ethical framework while preserving general competence. We present Convergent-Divergent Routing, which traces and edits minimal branch points inside transformer blocks where ethical-framework-related pathways first converge and then diverge. Gating non-target branches at these loci blocks the downstream propagation while leaving upstream computations intact. We find that this intervention alone increases targeted ethical-framework reasoning. To achieve fine-grained control, we adapt Common Spatial Patterns to the residual stream and extract, for each branch-point layer, a pair of directions that discriminate between utilitarian and deontological frameworks. We then introduce Dual Logit Calibration, a closed-form, minimum-\ell₂-norm update that moves the residual within this two-dimensional subspace so the resulting directional projections align with user-specified preference weights. Experiments on real-life moral dilemmas show that our method reliably achieves preference calibration and largely preserves general capabilities, outperforming recent baselines while providing an interpretable mechanism.
%U https://aclanthology.org/2026.acl-long.1933/
%P 41698-41721

Markdown (Informal)

[Where Paths Split: Localized, Calibrated Control of Moral Reasoning in Large Language Models](https://aclanthology.org/2026.acl-long.1933/) (Yuan et al., ACL 2026)

• Where Paths Split: Localized, Calibrated Control of Moral Reasoning in Large Language Models (Yuan et al., ACL 2026)

ACL

• Chenchen Yuan, Zheyu Zhang, and Gjergji Kasneci. 2026. Where Paths Split: Localized, Calibrated Control of Moral Reasoning in Large Language Models. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 41698–41721, San Diego, California, United States. Association for Computational Linguistics.