Dummy Backdoor as a Defense: Removing Unknown Backdoors via Shared Internal Mechanisms for Generative LLMs

Computer Science > Cryptography and Security [Submitted on 10 Jun 2026] Dummy Backdoor as a Defense: Removing Unknown Backdoors via Shared Internal Mechanisms for Generative LLMs Kazuki Iwahana, Masaru Matsubayashi, Takuma Koyama, Toshiki Shibahara, Kenichiro Omintato, Akira Ito Backdoor attacks pose a serious threat to the safety and reliability of Large Language Models (LLMs), as they cause models to behave normally on clean inputs while producing attacker-specified responses when hidden triggers are present. Removing such unknown backdoors is particularly challenging when the defender does not know the backdoor attack types or the internal mechanisms formed through backdoor training. In this work, we propose a simple but effective backdoor removal method based on shared internal mechanisms across different backdoors. First, we show that different backdoors with the same task (attack objective) induce similar trigger-activated changes in the internal activations. Motivated by this observation, our method intentionally embeds a backdoor with a known trigger (\emph{dummy backdoor}) and then removes it through further fine-tuning on dummy-triggered inputs paired with clean responses. Since the dummy backdoor and the unknown backdoor can rely on shared internal mechanisms, removing the dummy backdoor also reduces the effect of the unknown backdoor. We evaluate our method on three backdoor attack types across multiple model families. Experimental results show that our method substantially reduces the attack success rate of the unknown backdoor while preserving model utility, outperforming representative existing defense methods in both backdoor removal effectiveness and utility preservation. These findings suggest that a defender-controllable backdoor can serve as a helpful proxy for mitigating unknown backdoors in generative LLMs. Subjects: Cryptography and Security (cs.CR); Computation and Language (cs.CL) Cite as: arXiv:2606.11648 [cs.CR]   (or arXiv:2606.11648v1 [cs.CR] for this version)   https://doi.org/10.48550/arXiv.2606.11648 Focus to learn more Submission history From: Kazuki Iwahana [view email] [v1] Wed, 10 Jun 2026 04:26:57 UTC (376 KB) Access Paper: HTML (experimental) view license Current browse context: cs.CR < prev   |   next > new | recent | 2026-06 Change to browse by: cs cs.CL References & Citations NASA ADS Google Scholar Semantic Scholar Export BibTeX Citation Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Demos Related Papers About arXivLabs Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)