Unlearning Backdoor Attacks for LLMs with Weak-to-Strong Knowledge Distillation
Parameter-efficient fine-tuning (PEFT) can bridge the gap between large language models (LLMs) and downstream tasks. However, PEFT has been proven vulnerable to malicious attacks. Research indicates that poisoned LLMs, even after PEFT, retain the capability to activate internalized backdoors when input samples contain predefined triggers. In this paper, we introduce a novel weak-to-strong unlearning algorithm to defend against backdoor attacks based on feature alignment knowledge distillation, named W2SDefense. Specifically, we first train a small-scale language model through full-parameter fine-tuning to serve as the clean teacher model. Then, this teacher model guides the large-scale poisoned student model in unlearning the backdoor, leveraging PEFT. Theoretical analysis suggests that W2SDefense has the potential to enhance the student model's ability to unlearn backdoor features, preventing the activation of the backdoor. We conduct comprehensive experiments on three state-of-the-art large language models and several different backdoor attack algorithms. Our empirical results demonstrate the outstanding performance of W2SDefense in defending against backdoor attacks without compromising model performance.
Code (1)
Tasks
Backdoor AttackKnowledge DistillationLanguage ModelingLanguage Modellingparameter-efficient fine-tuningtext-classificationText ClassificationSimilar Papers 제목 키워드 기반
Forgetting to Forget: Attention Sink as A Gateway for Backdooring LLM Unlearning
Large language model (LLM) unlearning is a key approach for removing undesired data, knowledge, or behaviors from pretrained models while retaining their general utility. Yet, with the rise of open-weight LLMs, we ask: c…
Injection, Attack and Erasure: Revocable Backdoor Attacks via Machine Unlearning
Backdoor attacks pose a persistent security risk to deep neural networks (DNNs) due to their stealth and durability. While recent research has explored leveraging model unlearning mechanisms to enhance backdoor concealme…
Bilevel OptimizationWeak-to-Strong Backdoor Attack for Large Language Models
Despite being widely applied due to their exceptional capabilities, Large Language Models (LLMs) have been proven to be vulnerable to backdoor attacks. These attacks introduce targeted vulnerabilities into LLMs by poison…
Backdoor AttackKnowledge Distillationparameter-efficient fine-tuningBackdoor Unlearning Generalization: A Path Toward the Removal of Unknown Triggers in LLMs
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…
Continual PretrainingMitigating Backdoor Attacks using Activation-Guided Model Editing
Backdoor attacks compromise the integrity and reliability of machine learning models by embedding a hidden trigger during the training process, which can later be activated to cause unintended misbehavior. We propose a n…
Machine UnlearningModel Editing