The paper introduces a low‑rank auditing method called LoRA as Oracle, which fits a small adapter to a hypothesis and analyzes the geometry, energy, and alignment of the resulting update relative to frozen weights. This approach directly measures what a model has internalized, independent of its output behavior, enabling detection of backdoors that behavioral audits miss. By identifying and erasing malicious internalizations within the same low‑rank subspace, the method consistently detects target classes across multiple datasets and architectures while preserving clean accuracy and operating at far lower parameter and memory cost than full‑model baselines.
arXiv:2607. 26849v1 Announce Type: cross Abstract: As large language models (LLMs) are deployed in high-stakes domains, adversaries may poison training data to implant backdoors: hidden triggers that covertly manipulate model behavior at inference time.
As large language models (LLMs) are deployed in high-stakes domains, adversaries may poison training data to implant backdoors: hidden triggers that covertly manipulate model behavior at inference time. We ask whether a defender can recover such a trigger under realistic affordances, namely white-box access to the weights and knowledge of the behavior of concern, but no training data, no trusted reference model, no knowledge of the trigger, and no certainty that the model is poisoned.
arXiv:2608. 06795v1 Announce Type: cross Abstract: Low-rank adaptation (LoRA) enables efficient specialization and distribution of large language models through compact adapters.
arXiv:2608.24354v1 Announce Type: cross Abstract: MLLMs are increasingly deployed in user-facing applications, yet they inherit backdoor risks from the pipelines used to construct them: triggers may...
The paper introduces a method to purify LoRA-tuned large language models (LLMs) against backdoor attacks without relying on trigger knowledge, clean references, or retraining. By extracting high‑fidelity backdoor directions and projecting LoRA updates onto orthogonal null spaces in input and output channels, the approach reduces attack success rates from nearly 100% to under 10%. Experiments demonstrate that this null‑space projection preserves both the base model’s general capabilities and the new downstream skills learned through the adapter across various tasks.
arXiv:2607. 19894v1 Announce Type: cross Abstract: Large language models (LLMs) are vulnerable to backdoor attacks, where hidden triggers induce malicious outputs.
arXiv:2508. 04064v2 Announce Type: replace-cross Abstract: Horizontal federated learning (HFL) backdoor audits often summarize model behavior through clean accuracy (CA), mean attack success rate (ASR), or a single known-trigger test.
arXiv:2510.17021v2 Announce Type: replace-cross Abstract: Large language model (LLM) unlearning is a key approach for removing undesired data, knowledge, or behaviors from pretrained models while ret...
Large language models (LLMs) are vulnerable to backdoor attacks, where hidden triggers induce malicious outputs. Existing defenses generally fall into inference-time detection or training-time mitigation, but face two key limitations.
arXiv:2507. 05113v3 Announce Type: replace-cross Abstract: Deep Neural Networks (DNNs) are susceptible to backdoor attacks, where adversaries poison training data to implant backdoor into the victim model.
arXiv:2608. 00732v1 Announce Type: new Abstract: Backdoor attacks pose a serious threat to deep neural networks, especially when training relies on third-party data, allowing adversaries to inject malicious behaviors through data poisoning.
The paper introduces TRIM, a black‑box defense for backdoor attacks in computer vision models. TRIM identifies and removes malicious trigger regions at inference time using region‑based segmentation, adaptive trigger discovery via inpainting and diffusion, and selective purification, without needing model internals, training data, or clean samples. Experiments on various datasets and trigger types show TRIM reduces attack success rates to as low as 1.16% while maintaining high clean accuracy.