@inproceedings{garrido-munoz-etal-2026-towards,
title = "Towards Reliable {AI} Fairness: Challenges in Steering Features within Bias-Implicated Neurons",
author = "Garrido-Munoz, Ismael and
Montejo-Raez, Arturo and
Mart{\'i}nez-Santiago, Fernando",
editor = "Piperidis, Stelios and
Bel, N{\'u}ria and
van den Heuvel, Henk and
Ide, Nancy and
Krek, Simon and
Toral, Antonio",
booktitle = "Proceedings of the Fifteenth Language Resources and Evaluation Conference",
month = may,
year = "2026",
address = "Palma de Mallorca, Spain",
publisher = "ELRA Language Resource Association",
url = "https://aclanthology.org/2026.lrec-1.306/",
doi = "10.63317/2iexsnkqn3j6",
pages = "3851--3860",
abstract = "LLMs perpetuate societal biases, such as gender stereotypes, reinforcing harmful norms and posing significant fairness risks in real-world applications. We investigate a fine-grained mitigation technique that moves beyond surface-level fixes. Our approach uses attribution graphs to identify and directly steer bias-implicated features within a Sparse Autoencoder{'}s (SAE) latent space. This method, known as feature steering, offers a theoretically precise, surgical intervention aimed at correcting bias at its neural source without costly retraining. We critically examine its practical reliability across various contexts. We find that steering effectiveness is highly sensitive to parameter tuning, often requiring unpredictable, context-specific adjustments. The intervention{'}s success exists in narrow ``sweet spots,'' outside of which performance can degrade catastrophically. This demonstrates that while direct intervention on learned features is a powerful analytical tool, significant challenges of brittleness and instability hinder its application as a consistent, broad-scale debiasing solution, necessitating research into more robust control mechanisms."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="garrido-munoz-etal-2026-towards">
<titleInfo>
<title>Towards Reliable AI Fairness: Challenges in Steering Features within Bias-Implicated Neurons</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ismael</namePart>
<namePart type="family">Garrido-Munoz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arturo</namePart>
<namePart type="family">Montejo-Raez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fernando</namePart>
<namePart type="family">Martínez-Santiago</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Fifteenth Language Resources and Evaluation Conference</title>
</titleInfo>
<name type="personal">
<namePart type="given">Stelios</namePart>
<namePart type="family">Piperidis</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Núria</namePart>
<namePart type="family">Bel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Henk</namePart>
<namePart type="family">van den Heuvel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nancy</namePart>
<namePart type="family">Ide</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Simon</namePart>
<namePart type="family">Krek</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Antonio</namePart>
<namePart type="family">Toral</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resource Association</publisher>
<place>
<placeTerm type="text">Palma de Mallorca, Spain</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>LLMs perpetuate societal biases, such as gender stereotypes, reinforcing harmful norms and posing significant fairness risks in real-world applications. We investigate a fine-grained mitigation technique that moves beyond surface-level fixes. Our approach uses attribution graphs to identify and directly steer bias-implicated features within a Sparse Autoencoder’s (SAE) latent space. This method, known as feature steering, offers a theoretically precise, surgical intervention aimed at correcting bias at its neural source without costly retraining. We critically examine its practical reliability across various contexts. We find that steering effectiveness is highly sensitive to parameter tuning, often requiring unpredictable, context-specific adjustments. The intervention’s success exists in narrow “sweet spots,” outside of which performance can degrade catastrophically. This demonstrates that while direct intervention on learned features is a powerful analytical tool, significant challenges of brittleness and instability hinder its application as a consistent, broad-scale debiasing solution, necessitating research into more robust control mechanisms.</abstract>
<identifier type="citekey">garrido-munoz-etal-2026-towards</identifier>
<identifier type="doi">10.63317/2iexsnkqn3j6</identifier>
<location>
<url>https://aclanthology.org/2026.lrec-1.306/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>3851</start>
<end>3860</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Towards Reliable AI Fairness: Challenges in Steering Features within Bias-Implicated Neurons
%A Garrido-Munoz, Ismael
%A Montejo-Raez, Arturo
%A Martínez-Santiago, Fernando
%Y Piperidis, Stelios
%Y Bel, Núria
%Y van den Heuvel, Henk
%Y Ide, Nancy
%Y Krek, Simon
%Y Toral, Antonio
%S Proceedings of the Fifteenth Language Resources and Evaluation Conference
%D 2026
%8 May
%I ELRA Language Resource Association
%C Palma de Mallorca, Spain
%F garrido-munoz-etal-2026-towards
%X LLMs perpetuate societal biases, such as gender stereotypes, reinforcing harmful norms and posing significant fairness risks in real-world applications. We investigate a fine-grained mitigation technique that moves beyond surface-level fixes. Our approach uses attribution graphs to identify and directly steer bias-implicated features within a Sparse Autoencoder’s (SAE) latent space. This method, known as feature steering, offers a theoretically precise, surgical intervention aimed at correcting bias at its neural source without costly retraining. We critically examine its practical reliability across various contexts. We find that steering effectiveness is highly sensitive to parameter tuning, often requiring unpredictable, context-specific adjustments. The intervention’s success exists in narrow “sweet spots,” outside of which performance can degrade catastrophically. This demonstrates that while direct intervention on learned features is a powerful analytical tool, significant challenges of brittleness and instability hinder its application as a consistent, broad-scale debiasing solution, necessitating research into more robust control mechanisms.
%R 10.63317/2iexsnkqn3j6
%U https://aclanthology.org/2026.lrec-1.306/
%U https://doi.org/10.63317/2iexsnkqn3j6
%P 3851-3860
Markdown (Informal)
[Towards Reliable AI Fairness: Challenges in Steering Features within Bias-Implicated Neurons](https://aclanthology.org/2026.lrec-1.306/) (Garrido-Munoz et al., LREC 2026)
ACL