@inproceedings{guo-etal-2025-elephant,
title = "The Elephant in the Room: Exploring the Role of Neutral Words in Language Model Group-Agnostic Debiasing",
author = "Guo, Xinwei and
Gao, Jiashi and
Zhou, Junlei and
Zhang, Jiaxin and
Chen, Guanhua and
Zhao, Xiangyu and
Liu, Quanying and
Wu, Haiyan and
Yao, Xin and
Wei, Xuetao",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-acl.1044/",
doi = "10.18653/v1/2025.findings-acl.1044",
pages = "20360--20371",
ISBN = "979-8-89176-256-5",
abstract = "Large Language Models (LLMs) are increasingly integrated into our daily lives, raising significant ethical concerns, especially about perpetuating stereotypes.While group-specific debiasing methods have made progress, they often fail to address multiple biases simultaneously. In contrast, group-agnostic debiasing has the potential to mitigate a variety of biases at once, but remains underexplored.In this work, we investigate the role of neutral words{---}the group-agnostic component{---}in enhancing the group-agnostic debiasing process. We first reveal that neutral words are essential for preserving semantic modeling, and we propose $\epsilon$-DPCE, a method that incorporates a neutral word semantics-based loss function to effectively alleviate the deterioration of the Language Modeling Score (LMS) during the debiasing process. Furthermore, by introducing the SCM-Projection method, we demonstrate that SCM-based debiasing eliminates stereotypes by indirectly disrupting the association between attribute and neutral words in the Stereotype Content Model (SCM) space. Our experiments show that neutral words, which often embed multi-group stereotypical objects, play a key role in contributing to the group-agnostic nature of SCM-based debiasing."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="guo-etal-2025-elephant">
<titleInfo>
<title>The Elephant in the Room: Exploring the Role of Neutral Words in Language Model Group-Agnostic Debiasing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xinwei</namePart>
<namePart type="family">Guo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiashi</namePart>
<namePart type="family">Gao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junlei</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiaxin</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guanhua</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiangyu</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Quanying</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haiyan</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xin</namePart>
<namePart type="family">Yao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xuetao</namePart>
<namePart type="family">Wei</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-256-5</identifier>
</relatedItem>
<abstract>Large Language Models (LLMs) are increasingly integrated into our daily lives, raising significant ethical concerns, especially about perpetuating stereotypes.While group-specific debiasing methods have made progress, they often fail to address multiple biases simultaneously. In contrast, group-agnostic debiasing has the potential to mitigate a variety of biases at once, but remains underexplored.In this work, we investigate the role of neutral words—the group-agnostic component—in enhancing the group-agnostic debiasing process. We first reveal that neutral words are essential for preserving semantic modeling, and we propose ε-DPCE, a method that incorporates a neutral word semantics-based loss function to effectively alleviate the deterioration of the Language Modeling Score (LMS) during the debiasing process. Furthermore, by introducing the SCM-Projection method, we demonstrate that SCM-based debiasing eliminates stereotypes by indirectly disrupting the association between attribute and neutral words in the Stereotype Content Model (SCM) space. Our experiments show that neutral words, which often embed multi-group stereotypical objects, play a key role in contributing to the group-agnostic nature of SCM-based debiasing.</abstract>
<identifier type="citekey">guo-etal-2025-elephant</identifier>
<identifier type="doi">10.18653/v1/2025.findings-acl.1044</identifier>
<location>
<url>https://aclanthology.org/2025.findings-acl.1044/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>20360</start>
<end>20371</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T The Elephant in the Room: Exploring the Role of Neutral Words in Language Model Group-Agnostic Debiasing
%A Guo, Xinwei
%A Gao, Jiashi
%A Zhou, Junlei
%A Zhang, Jiaxin
%A Chen, Guanhua
%A Zhao, Xiangyu
%A Liu, Quanying
%A Wu, Haiyan
%A Yao, Xin
%A Wei, Xuetao
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Findings of the Association for Computational Linguistics: ACL 2025
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-256-5
%F guo-etal-2025-elephant
%X Large Language Models (LLMs) are increasingly integrated into our daily lives, raising significant ethical concerns, especially about perpetuating stereotypes.While group-specific debiasing methods have made progress, they often fail to address multiple biases simultaneously. In contrast, group-agnostic debiasing has the potential to mitigate a variety of biases at once, but remains underexplored.In this work, we investigate the role of neutral words—the group-agnostic component—in enhancing the group-agnostic debiasing process. We first reveal that neutral words are essential for preserving semantic modeling, and we propose ε-DPCE, a method that incorporates a neutral word semantics-based loss function to effectively alleviate the deterioration of the Language Modeling Score (LMS) during the debiasing process. Furthermore, by introducing the SCM-Projection method, we demonstrate that SCM-based debiasing eliminates stereotypes by indirectly disrupting the association between attribute and neutral words in the Stereotype Content Model (SCM) space. Our experiments show that neutral words, which often embed multi-group stereotypical objects, play a key role in contributing to the group-agnostic nature of SCM-based debiasing.
%R 10.18653/v1/2025.findings-acl.1044
%U https://aclanthology.org/2025.findings-acl.1044/
%U https://doi.org/10.18653/v1/2025.findings-acl.1044
%P 20360-20371
Markdown (Informal)
[The Elephant in the Room: Exploring the Role of Neutral Words in Language Model Group-Agnostic Debiasing](https://aclanthology.org/2025.findings-acl.1044/) (Guo et al., Findings 2025)
ACL
- Xinwei Guo, Jiashi Gao, Junlei Zhou, Jiaxin Zhang, Guanhua Chen, Xiangyu Zhao, Quanying Liu, Haiyan Wu, Xin Yao, and Xuetao Wei. 2025. The Elephant in the Room: Exploring the Role of Neutral Words in Language Model Group-Agnostic Debiasing. In Findings of the Association for Computational Linguistics: ACL 2025, pages 20360–20371, Vienna, Austria. Association for Computational Linguistics.