Architectural Backdoors in Vision-Language Model Supply Chains via Representation Steering

• URL: http://arxiv.org/abs/2607.25479v1
• Date: Tue, 28 Jul 2026 09:12:11 GMT
• Title: Architectural Backdoors in Vision-Language Model Supply Chains via Representation Steering
• Abstract summary: Vision--Language Models (VLMs) are increasingly deployed through a model supply chain.<n>In this paper, we show that a malicious provider can exploit this trust boundary by embedding architectural backdoors into VLM supply chains.
• Score: 9.737607721974664
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Vision--Language Models (VLMs) are increasingly deployed through a model supply chain in which pretrained checkpoints, architecture definitions, text encoders, and exported computation graphs are distributed by third parties and reused across downstream services. This reuse model creates a security-critical trust boundary: VLM deployments inherit not only learned parameters but also executable behavior encoded in shared model artifacts. In this paper, we show that a malicious provider can exploit this trust boundary by embedding architectural backdoors into VLM supply chains through representation steering. Our attack introduces dormant steering logic into the model architecture through a trigger-gated additive modification of an intermediate representation, without poisoning training data, controlling downstream fine-tuning, or modifying prompts at deployment time. When the trigger is absent, the modification reduces to zero and the model follows its normal computation, preserving clean utility. When the trigger is present, a steering direction shifts the internal representation toward an attacker-defined objective. We evaluate the attack across multiple VLM families and downstream tasks, including visual question answering, text-to-image generation, retrieval, and semantic response biasing. The results show that the proposed architectural steering backdoor compromises integrity, safety enforcement, and ranking fairness while preserving normal behavior on clean inputs. We further show that shared VLM artifacts can carry dormant steering logic against downstream services, and we propose an auditing defense that inspects the executable logic distributed with model artifacts rather than only their learned weights.

Related papers

• Defending against Model Extraction for GNNs with Model Reprogramming [57.82196268649236]
  Graph Neural Networks (GNNs) serve as the backbone for high-stakes applications in Machine-Learning-as-a-Service (ML)<n>Still, their black-box deployment exposes them to Model Extraction (ME) attacks.<n>Existing defenses suffer from a critical ''Euclidean bias''<n>We propose GraphRP, a proactive defense framework that repurposes Model Reprogramming for security.
  arXiv  Detail & Related papers  (2026-08-11T23:20:15Z)
• Beyond the Prompt: Jailbreaking Function-Calling LLMs via Simulated Moderation Traces [32.850551109913845]
  Jailbreak attacks remain a critical threat to the safe deployment of large language models (LLMs)<n>We show that this prompt-centric paradigm overlooks a structural vulnerability in stateful, function-calling environments.<n>We exploit this architectural flaw through SMT, a black-box attack framework based on Simulated Moderation Traces.
  arXiv  Detail & Related papers  (2026-07-01T06:08:07Z)
• Beyond Homophily: Towards Generalized Graph Reconstruction Attack and Defense [80.28330585803133]
  Graph neural networks (GNNs) are widely deployed on computation data, yet they can leak sensitive or proprietary information about the training graph adjacency.<n>This work studies graph reconstruction attacks (GRA), a form of model inversion that reconstructs the training adjacency from a trained GNN.<n>We propose MC-GRA (+), which reconstructs the adjacency by optimizing a surrogate adjacency whose GNN-induced representations align with those of the target model.<n>On the defense side, we propose MC-GPB (+), which suppresses adjacency-dependent information throughout the representation chain
  arXiv  Detail & Related papers  (2026-06-06T09:29:27Z)
• Attesting LLM Pipelines: Enforcing Verifiable Training and Release Claims [2.0403981727850233]
  Modern Large Language Model (LLM) systems are assembled from third-party artifacts such as pre-trained weights, fine-tuning adapters, datasets, dependency packages, and container images.<n>This speed comes with supply-chain risks, including compromised dependencies, malicious hub artifacts, unsafe dedefaultization, forged provenance, and backdoored models.<n>We propose an attestation-aware promotion gate: before an artifact is admitted into trusted environments, the gate verifies claim evidence, enforces safe loading and static scanning policies, and applies secure-by-serial deployment constraints.
  arXiv  Detail & Related papers  (2026-03-30T20:37:48Z)
• Shape and Substance: Dual-Layer Side-Channel Attacks on Local Vision-Language Models [2.1198879079315573]
  On-device Vision-Language Models (VLMs) promise data privacy via local execution.<n>We show that the architectural shift toward Dynamic High-Resolution preprocessing introduces an inherent algorithmic side-channel.<n>We demonstrate a dual-layer attack framework against local VLMs.
  arXiv  Detail & Related papers  (2026-03-26T12:53:49Z)
• Backdoor Directions in Vision Transformers [56.382912038371046]
  This paper investigates how Backdoor Attacks are represented within Vision Transformers (ViTs)<n>We identify a specific trigger direction'' in the model's activations that corresponds to the internal representation of the trigger.<n>Using this direction as a diagnostic tool, we trace how backdoor features are processed across layers.
  arXiv  Detail & Related papers  (2026-03-11T14:13:48Z)
• CREDIT: Certified Ownership Verification of Deep Neural Networks Against Model Extraction Attacks [54.04030169323115]
  We introduce CREDIT, a certified ownership verification against Model Extraction Attacks (MEAs)<n>We quantify the similarity between DNN models, propose a practical verification threshold, and provide rigorous theoretical guarantees for ownership verification based on this threshold.<n>We extensively evaluate our approach on several mainstream datasets across different domains and tasks, achieving state-of-the-art performance.
  arXiv  Detail & Related papers  (2026-02-23T23:36:25Z)
• Efficient Representations are Controllable Representations [0.0]
  Controlling how LLMs internally represent concepts requires sophisticated methods to first identify, then intervene on the model's existing feature geometry.<n>We finetune an LLM with a simple auxiliary loss, training 16 of its 3072 residual stream dimensions to be inert interpretability flags that simply indicate what concepts are required for generation.<n>As a result, these inert flags become genuine internal features: interpretable control switches that allow us to steer generation at inference time.
  arXiv  Detail & Related papers  (2026-02-08T05:32:02Z)
• Every Step Counts: Decoding Trajectories as Authorship Fingerprints of dLLMs [63.82840470917859]
  We show that the decoding mechanism of dLLMs can be used as a powerful tool for model attribution.<n>We propose a novel information extraction scheme called the Directed Decoding Map (DDM), which captures structural relationships between decoding steps and better reveals model-specific behaviors.
  arXiv  Detail & Related papers  (2025-10-02T06:25:10Z)
• Proactive Disentangled Modeling of Trigger-Object Pairings for Backdoor Defense [0.0]
  Deep neural networks (DNNs) and generative AI (GenAI) are increasingly vulnerable to backdoor attacks.<n>In this paper, we introduce DBOM, a proactive framework that leverages structured disentanglement to identify and neutralize both seen and unseen backdoor threats.<n>We show that DBOM robustly detects poisoned images prior to downstream training, significantly enhancing the security of training pipelines.
  arXiv  Detail & Related papers  (2025-08-03T21:58:15Z)
• CogCoM: A Visual Language Model with Chain-of-Manipulations Reasoning [61.21923643289266]
  Chain of Manipulations is a mechanism that enables Vision-Language Models to solve problems step-by-step with evidence.<n>After training, models can solve various visual problems by eliciting intrinsic manipulations (e.g., grounding, zoom in) actively without involving external tools.<n>Our trained model, textbfCogCoM, achieves state-of-the-art performance across 9 benchmarks from 4 categories.
  arXiv  Detail & Related papers  (2024-02-06T18:43:48Z)
• Defending Variational Autoencoders from Adversarial Attacks with MCMC [74.36233246536459]
  Variational autoencoders (VAEs) are deep generative models used in various domains. As previous work has shown, one can easily fool VAEs to produce unexpected latent representations and reconstructions for a visually slightly modified input. Here, we examine several objective functions for adversarial attacks construction, suggest metrics assess the model robustness, and propose a solution.
  arXiv  Detail & Related papers  (2022-03-18T13:25:18Z)

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.