@inproceedings{lin-etal-2026-backdoor,
title = "Backdoor Collapse: Eliminating Unknown Threats Via Known Backdoor Aggregation In Language Models",
author = "Lin, Liang and
Yu, Miao and
Aloqaily, Moayad and
Zhou, Zhenhong and
Wang, Kun and
Pang, Linsey and
Mehrotra, Prakhar and
Wen, Qingsong",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.acl-long.920/",
pages = "20105--20123",
ISBN = "979-8-89176-390-6",
abstract = "Backdoor attacks are a significant threat to large language models (LLMs), often embedded via public checkpoints, yet existing defenses rely on impractical assumptions about trigger settings. To address this challenge, we propose Locphylax, a defense framework that requires no prior knowledge of trigger settings. Locphylax is based on the key observation that when deliberately injecting known backdoors into an already-compromised model, both existing unknown and newly injected backdoors aggregate in the representation space. Locphylax leverages this through a two-stage process: first, aggregating backdoor representations by injecting known triggers, and then, performing recovery fine-tuning to restore benign outputs. Extensive experiments across multiple LLM architectures demonstrate that: (I) Locphylax reduces the average Attack Success Rate to 4.41{\%} across multiple benchmarks, outperforming existing baselines by 28.1{\%}{--}69.3{\%}. (II) Clean accuracy and utility are preserved within 0.5{\%} of the original model, ensuring negligible impact on legitimate tasks. (III) The defense generalizes across different types of backdoors, confirming its robustness in practical deployment scenarios. Our code is available at https://anonymous.4open.science/r/Locphylax."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lin-etal-2026-backdoor">
<titleInfo>
<title>Backdoor Collapse: Eliminating Unknown Threats Via Known Backdoor Aggregation In Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Liang</namePart>
<namePart type="family">Lin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Miao</namePart>
<namePart type="family">Yu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Moayad</namePart>
<namePart type="family">Aloqaily</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhenhong</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kun</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Linsey</namePart>
<namePart type="family">Pang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Prakhar</namePart>
<namePart type="family">Mehrotra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qingsong</namePart>
<namePart type="family">Wen</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>Backdoor attacks are a significant threat to large language models (LLMs), often embedded via public checkpoints, yet existing defenses rely on impractical assumptions about trigger settings. To address this challenge, we propose Locphylax, a defense framework that requires no prior knowledge of trigger settings. Locphylax is based on the key observation that when deliberately injecting known backdoors into an already-compromised model, both existing unknown and newly injected backdoors aggregate in the representation space. Locphylax leverages this through a two-stage process: first, aggregating backdoor representations by injecting known triggers, and then, performing recovery fine-tuning to restore benign outputs. Extensive experiments across multiple LLM architectures demonstrate that: (I) Locphylax reduces the average Attack Success Rate to 4.41% across multiple benchmarks, outperforming existing baselines by 28.1%–69.3%. (II) Clean accuracy and utility are preserved within 0.5% of the original model, ensuring negligible impact on legitimate tasks. (III) The defense generalizes across different types of backdoors, confirming its robustness in practical deployment scenarios. Our code is available at https://anonymous.4open.science/r/Locphylax.</abstract>
<identifier type="citekey">lin-etal-2026-backdoor</identifier>
<location>
<url>https://aclanthology.org/2026.acl-long.920/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>20105</start>
<end>20123</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Backdoor Collapse: Eliminating Unknown Threats Via Known Backdoor Aggregation In Language Models
%A Lin, Liang
%A Yu, Miao
%A Aloqaily, Moayad
%A Zhou, Zhenhong
%A Wang, Kun
%A Pang, Linsey
%A Mehrotra, Prakhar
%A Wen, Qingsong
%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 lin-etal-2026-backdoor
%X Backdoor attacks are a significant threat to large language models (LLMs), often embedded via public checkpoints, yet existing defenses rely on impractical assumptions about trigger settings. To address this challenge, we propose Locphylax, a defense framework that requires no prior knowledge of trigger settings. Locphylax is based on the key observation that when deliberately injecting known backdoors into an already-compromised model, both existing unknown and newly injected backdoors aggregate in the representation space. Locphylax leverages this through a two-stage process: first, aggregating backdoor representations by injecting known triggers, and then, performing recovery fine-tuning to restore benign outputs. Extensive experiments across multiple LLM architectures demonstrate that: (I) Locphylax reduces the average Attack Success Rate to 4.41% across multiple benchmarks, outperforming existing baselines by 28.1%–69.3%. (II) Clean accuracy and utility are preserved within 0.5% of the original model, ensuring negligible impact on legitimate tasks. (III) The defense generalizes across different types of backdoors, confirming its robustness in practical deployment scenarios. Our code is available at https://anonymous.4open.science/r/Locphylax.
%U https://aclanthology.org/2026.acl-long.920/
%P 20105-20123
Markdown (Informal)
[Backdoor Collapse: Eliminating Unknown Threats Via Known Backdoor Aggregation In Language Models](https://aclanthology.org/2026.acl-long.920/) (Lin et al., ACL 2026)
ACL
- Liang Lin, Miao Yu, Moayad Aloqaily, Zhenhong Zhou, Kun Wang, Linsey Pang, Prakhar Mehrotra, and Qingsong Wen. 2026. Backdoor Collapse: Eliminating Unknown Threats Via Known Backdoor Aggregation In Language Models. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 20105–20123, San Diego, California, United States. Association for Computational Linguistics.