Backdoor Unlearning Generalization: A Path Toward the Removal of Unknown Triggers in LLMs

• URL: http://arxiv.org/abs/2606.03785v2
• Date: Thu, 04 Jun 2026 08:44:14 GMT
• Title: Backdoor Unlearning Generalization: A Path Toward the Removal of Unknown Triggers in LLMs
• Abstract summary: We show that backdoor neutralization through unlearning generalizes across backdoors.<n>We study this phenomenon across three model families, whose backdoors were injected via pretraining or continual pretraining.<n>Our results open a new direction for LLM safety as defenders could deliberately inject controlled backdoors and then remove them.
• Score: 2.855228082831898
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Backdoor attacks in Large Language Models (LLMs) are a growing security concern, where models can generate adversary-chosen content. Existing defenses target backdoors one at a time and typically require knowledge of the trigger, leaving the defender at a structural disadvantage when unknown backdoors may exist in a model. We show that backdoor neutralization through unlearning generalizes across backdoors: training a model to ignore a single trigger can also suppress other backdoors that were never explicitly targeted. We study this phenomenon across three model families, whose backdoors were injected via pretraining or continual pretraining, by analyzing the models obtained after removing one backdoor at a time. To understand why unlearning certain backdoors induces the suppression of others, we introduce the Cross Activation Shift Distance, to quantify the distance between model changes induced by different trainings. Our results open a new direction for LLM safety as defenders could deliberately inject controlled backdoors and then remove them, leveraging cross-backdoor transfer to also suppress unknown backdoors that an attacker may have previously introduced in the model.

Related papers

• Dummy Backdoor as a Defense: Removing Unknown Backdoors via Shared Internal Mechanisms for Generative LLMs [1.4363317131844815]
  Backdoor attacks pose a serious threat to the safety and reliability of Large Language Models.<n>We propose a simple but effective backdoor removal method based on shared internal mechanisms across different backdoors.<n>Our method substantially reduces the attack success rate of the unknown backdoor while preserving model utility.
  arXiv  Detail & Related papers  (2026-06-10T04:26:57Z)
• Expose Before You Defend: Unifying and Enhancing Backdoor Defenses via Exposed Models [68.40324627475499]
  We introduce a novel two-step defense framework named Expose Before You Defend. EBYD unifies existing backdoor defense methods into a comprehensive defense system with enhanced performance. We conduct extensive experiments on 10 image attacks and 6 text attacks across 2 vision datasets and 4 language datasets.
  arXiv  Detail & Related papers  (2024-10-25T09:36:04Z)
• Flatness-aware Sequential Learning Generates Resilient Backdoors [7.969181278996343]
  Recently, backdoor attacks have become an emerging threat to the security of machine learning models. This paper counters CF of backdoors by leveraging continual learning (CL) techniques. We propose a novel framework, named Sequential Backdoor Learning (SBL), that can generate resilient backdoors.
  arXiv  Detail & Related papers  (2024-07-20T03:30:05Z)
• Mitigating Backdoor Attack by Injecting Proactive Defensive Backdoor [63.84477483795964]
  Data-poisoning backdoor attacks are serious security threats to machine learning models. In this paper, we focus on in-training backdoor defense, aiming to train a clean model even when the dataset may be potentially poisoned. We propose a novel defense approach called PDB (Proactive Defensive Backdoor)
  arXiv  Detail & Related papers  (2024-05-25T07:52:26Z)
• BaDExpert: Extracting Backdoor Functionality for Accurate Backdoor Input Detection [42.021282816470794]
  We present a novel defense, against backdoor attacks on Deep Neural Networks (DNNs) Our defense falls within the category of post-development defenses that operate independently of how the model was generated. We show the feasibility of devising highly accurate backdoor input detectors that filter out the backdoor inputs during model inference.
  arXiv  Detail & Related papers  (2023-08-23T21:47:06Z)
• Universal Soldier: Using Universal Adversarial Perturbations for Detecting Backdoor Attacks [15.917794562400449]
  A deep learning model may be poisoned by training with backdoored data or by modifying inner network parameters. It is difficult to distinguish between clean and backdoored models without prior knowledge of the trigger. We propose a novel method called Universal Soldier for Backdoor detection (USB) and reverse engineering potential backdoor triggers via UAPs.
  arXiv  Detail & Related papers  (2023-02-01T20:47:58Z)
• Neurotoxin: Durable Backdoors in Federated Learning [73.82725064553827]
  federated learning systems have an inherent vulnerability during their training to adversarial backdoor attacks. We propose Neurotoxin, a simple one-line modification to existing backdoor attacks that acts by attacking parameters that are changed less in magnitude during training.
  arXiv  Detail & Related papers  (2022-06-12T16:52:52Z)
• Check Your Other Door! Establishing Backdoor Attacks in the Frequency Domain [80.24811082454367]
  We show the advantages of utilizing the frequency domain for establishing undetectable and powerful backdoor attacks. We also show two possible defences that succeed against frequency-based backdoor attacks and possible ways for the attacker to bypass them.
  arXiv  Detail & Related papers  (2021-09-12T12:44:52Z)
• Turn the Combination Lock: Learnable Textual Backdoor Attacks via Word Substitution [57.51117978504175]
  Recent studies show that neural natural language processing (NLP) models are vulnerable to backdoor attacks. Injected with backdoors, models perform normally on benign examples but produce attacker-specified predictions when the backdoor is activated. We present invisible backdoors that are activated by a learnable combination of word substitution.
  arXiv  Detail & Related papers  (2021-06-11T13:03:17Z)
• Black-box Detection of Backdoor Attacks with Limited Information and Data [56.0735480850555]
  We propose a black-box backdoor detection (B3D) method to identify backdoor attacks with only query access to the model. In addition to backdoor detection, we also propose a simple strategy for reliable predictions using the identified backdoored models.
  arXiv  Detail & Related papers  (2021-03-24T12:06:40Z)
• Attack of the Tails: Yes, You Really Can Backdoor Federated Learning [21.06925263586183]
  Federated Learning (FL) lends itself to adversarial attacks in the form of backdoors during training. An edge-case backdoor forces a model to misclassify on seemingly easy inputs that are however unlikely to be part of the training, or test data, i.e., they live on the tail of the input distribution. We show how these edge-case backdoors can lead to unsavory failures and may have serious repercussions on fairness.
  arXiv  Detail & Related papers  (2020-07-09T21:50:54Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.