Towards Merging-conditional Backdoor Unalignment against Large Language Models

Reported by arXiv cs.CR (Yukun Chen, Yidou Chen, Yiming Li, Xinyu Zhang, Baosong Yang, Hongwei Yao, Zhan Qin)

Summary by Cyber Security News14 Oct 2026, 06:00 CEST

About this summary. This is a short, independently written summary of an article first published by arXiv cs.CR. Cyber Security News did not report or verify the underlying story. Read the original: https://arxiv.org/abs/2610.14701

arXiv:2610.14701v1 Announce Type: new Abstract: Model merging has become widely adopted for enhancing the task-specific performance of Large Language Models (LLMs) without additional training. While the specialized model from third-party platforms may undergo safety audits (e.g., backdoor detection), we reveal a novel and critical threat: One-sided Merging-Conditional Backdoor (OMCB).

Read the full article at arXiv cs.CR →

Source attribution: headline and facts are from arXiv cs.CR (arxiv.org). Summary method: excerpt of the source description. See our source attribution policy.

Related coverage

Analyzing and Mitigating Poisoning Attacks on Flow-Based Intrusion Detection

Quiet Triggers, Loud Footprints: A Tri-Space Measurement Study of Backdoor Stealthiness

Hackers Backdoor Nextcloud Servers Through Critical Flaw in Collabora’s Built-in CODE App

Volexity Uncovers Stealthy Ivanti VPN Attack Using Three Zero-Day Vulnerabilities

When Does Trust Help? A Leakage-Free Evaluation of Representation- and Aggregation-Based Defenses Against Backdoors in Federated Intrusion Detection

Cancerous Backdoor Attack: Gradual Activation via Invisible Tokens and A Stage Controller