Exploiting the Vulnerability of Large Language Models via Defense-Aware Architectural Backdoor
This addresses a security problem for users of large language models by revealing a novel, persistent threat that is incremental but more robust than prior methods.
The paper tackles the vulnerability of large language models to a new type of backdoor attack that hides within the model architecture, demonstrating that this attack can survive fine-tuning and evade existing defenses like BDDR.
Deep neural networks (DNNs) have long been recognized as vulnerable to backdoor attacks. By providing poisoned training data in the fine-tuning process, the attacker can implant a backdoor into the victim model. This enables input samples meeting specific textual trigger patterns to be classified as target labels of the attacker's choice. While such black-box attacks have been well explored in both computer vision and natural language processing (NLP), backdoor attacks relying on white-box attack philosophy have hardly been thoroughly investigated. In this paper, we take the first step to introduce a new type of backdoor attack that conceals itself within the underlying model architecture. Specifically, we propose to design separate backdoor modules consisting of two functions: trigger detection and noise injection. The add-on modules of model architecture layers can detect the presence of input trigger tokens and modify layer weights using Gaussian noise to disturb the feature distribution of the baseline model. We conduct extensive experiments to evaluate our attack methods using two model architecture settings on five different large language datasets. We demonstrate that the training-free architectural backdoor on a large language model poses a genuine threat. Unlike the-state-of-art work, it can survive the rigorous fine-tuning and retraining process, as well as evade output probability-based defense methods (i.e. BDDR). All the code and data is available https://github.com/SiSL-URI/Arch_Backdoor_LLM.