<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yuan-vlachos-2026-evaluating">
    <titleInfo>
        <title>Evaluating the Semantic Specificity of Representation Steering in Language Models</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Moy</namePart>
        <namePart type="family">Yuan</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Andreas</namePart>
        <namePart type="family">Vlachos</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>Localized Representation Steering (LRS) is widely used to correct reasoning pathologies in large language models. However, standard benchmark evaluations can easily be fooled by superficial label overrides, creating a false impression of reasoning circuit repairs. In this work, we propose Cross-Rule Transfer (CRT), a diagnostic framework that audits representational interventions by evaluating them on rule families where the model is natively competent. Evaluating late-layer LRS for a widespread logical failure, contradiction blindness, reveals that the intervention merely injects a global label bias: applying the steering vector to rules the model already handles correctly (99.6% baseline) degrades performance to 40.4% by forcing false contradiction predictions. We support this diagnosis with four complementary controls (direct logit bias equivalence, control vector label-flipping, cross-model grafting, and early-layer steering checks), providing a rigorous methodology to distinguish genuine reasoning repairs from superficial label overrides.</abstract>
    <identifier type="citekey">yuan-vlachos-2026-evaluating</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.562/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>11853</start>
            <end>11871</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Evaluating the Semantic Specificity of Representation Steering in Language Models
%A Yuan, Moy
%A Vlachos, Andreas
%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 yuan-vlachos-2026-evaluating
%X Localized Representation Steering (LRS) is widely used to correct reasoning pathologies in large language models. However, standard benchmark evaluations can easily be fooled by superficial label overrides, creating a false impression of reasoning circuit repairs. In this work, we propose Cross-Rule Transfer (CRT), a diagnostic framework that audits representational interventions by evaluating them on rule families where the model is natively competent. Evaluating late-layer LRS for a widespread logical failure, contradiction blindness, reveals that the intervention merely injects a global label bias: applying the steering vector to rules the model already handles correctly (99.6% baseline) degrades performance to 40.4% by forcing false contradiction predictions. We support this diagnosis with four complementary controls (direct logit bias equivalence, control vector label-flipping, cross-model grafting, and early-layer steering checks), providing a rigorous methodology to distinguish genuine reasoning repairs from superficial label overrides.
%U https://aclanthology.org/2026.findings-emnlp.562/
%P 11853-11871

Markdown (Informal)

[Evaluating the Semantic Specificity of Representation Steering in Language Models](https://aclanthology.org/2026.findings-emnlp.562/) (Yuan & Vlachos, Findings 2026)

• Evaluating the Semantic Specificity of Representation Steering in Language Models (Yuan & Vlachos, Findings 2026)

ACL

• Moy Yuan and Andreas Vlachos. 2026. Evaluating the Semantic Specificity of Representation Steering in Language Models. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 11853–11871, Budapest, Hungary. Association for Computational Linguistics.