paper-with-me

Papers

The Trigger in the Haystack: Extracting and Reconstructing LLM Backdoor Triggers

2026-02-03 · Blake Bullwinkel, Giorgio Severi, Keegan Hines, Amanda Minnich, Ram Shankar Siva Kumar, Yonatan Zunger arxiv

Detecting whether a model has been poisoned is a longstanding problem in AI security. In this work, we present a practical scanner for identifying sleeper agent-style backdoors in causal language models. Our approach relies on two key findings: first, sleeper agents tend to memorize poisoning data, making it possible to leak backdoor examples using memory extraction techniques. Second, poisoned LLMs exhibit distinctive patterns in their output distributions and attention heads when backdoor triggers are present in the input. Guided by these observations, we develop a scalable backdoor scanning methodology that assumes no prior knowledge of the trigger or target behavior and requires only inference operations. Our scanner integrates naturally into broader defensive strategies and does not alter model performance. We show that our method recovers working triggers across multiple backdoor scenarios and a broad range of models and fine-tuning methods.

📄 PDF Abstract BibTeX arXiv:2602.03085

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Gungnir: Exploiting Stylistic Features in Images for Backdoor Attacks on Diffusion Models

2025-02-28 · Yu Pan, Jiahao Chen, Bingrong Dai, Lin Wang 외

In recent years, Diffusion Models (DMs) have demonstrated significant advances in the field of image generation. However, according to current research, DMs are vulnerable to backdoor attacks, which allow attackers to co…

Backdoor Attackbackdoor defenseImage Generation

Removing the Trigger, Not the Backdoor: Alternative Triggers and Latent Backdoors

2026-03-10 · Gorka Abad, Ermes Franch, Stefanos Koffas, Stjepan Picek arxiv

Current backdoor defenses assume that neutralizing a known trigger removes the backdoor. We show this trigger-centric view is incomplete: \emph{alternative triggers}, patterns perceptually distinct from training triggers…

DISTIL: Data-Free Inversion of Suspicious Trojan Inputs via Latent Diffusion

2025-07-30 · Hossein Mirzaei, Zeinab Taghavi, Sepehr Rezaee, Masoud Hadi 외 arxiv

Deep neural networks have demonstrated remarkable success across numerous tasks, yet they remain vulnerable to Trojan (backdoor) attacks, raising serious concerns about their safety in real-world mission-critical applica…

Object Detection

UNICORN: A Unified Backdoor Trigger Inversion Framework

2023-04-05 · Zhenting Wang, Kai Mei, Juan Zhai, Shiqing Ma

The backdoor attack, where the adversary uses inputs stamped with triggers (e.g., a patch) to activate pre-planted malicious behaviors, is a severe threat to Deep Neural Network (DNN) models. Trigger inversion is an effe…

Backdoor Attack

From Shortcuts to Triggers: Backdoor Defense with Denoised PoE

2023-05-24 · Qin Liu, Fei Wang, Chaowei Xiao, Muhao Chen

Language models are often at risk of diverse backdoor attacks, especially data poisoning. Thus, it is important to investigate defense solutions for addressing them. Existing backdoor defense methods mainly focus on back…

backdoor defenseData PoisoningDenoisingSentence+1