Brand-as-Memory: Vision-Language Models Encode Causal, Mechanistically Localizable Credibility Priors for News Sources

• URL: http://arxiv.org/abs/2607.03365v1
• Date: Fri, 03 Jul 2026 14:22:05 GMT
• Title: Brand-as-Memory: Vision-Language Models Encode Causal, Mechanistically Localizable Credibility Priors for News Sources
• Abstract summary: We show that vision-language models (VLMs) increasingly read news and web content as images, where the publisher's identity is visually present.<n>We introduce Cue, a cross-model diagnostic that measures which surface source cue overrides an article's content evidence.
• Score: 3.168307501254055
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Vision-language models (VLMs) increasingly read news and web content as images, where the publisher's identity is visually present. We show that VLMs carry a strong source-credibility prior keyed on outlet identity, and study it along three axes. (i) Cross-model benchmark. We introduce CueTrust, a cross-model diagnostic that measures which surface source cue overrides an article's content evidence via a Source-Override Index (SOI). Across seven VLMs and five cues, the vulnerability profile is model- and scale-dependent, and the override is outlet-identity-specific and encoding-invariant, firing from the masthead name, the logo image, or the bare domain, but not from a named author, in-text authority, or page layout (clean negative controls). (ii) Mechanistic account. For the brand cue, we give a full mechanistic account: swapping only the masthead moves credibility across an approximately 11 log-odds range that tracks professional ratings (rho = 0.88 with Media Bias/Fact Check). The prior is dual-coded (name and logo), strengthens with scale, is causally formed at layers 19-21, carried by interpretable seed-stable sparse-autoencoder features, and recurs at the same relative locus in a second model family. It overrides content (about 1.8x) as a signal-magnitude effect within a shared pathway, not a privileged route. Steering the localized direction selectively reduces the override (41% reduction) and generalizes to held-out outlets, confirming the prior is causally used, not merely decodable. Deployed VLMs may thus defer to source identity over the evidence in front of them, a reliability failure we can measure across models, localize, and causally probe. We release the stimulus suite and CueTrust.

Related papers

• Recognition Without Enforcement: Configuration-Dependent Failures in LLM Agent Instruction Arbitration and External Control [0.0]
  LLM agents arbitrate among instructions from system prompts, users, memory, and tools, but this arbitration cannot be assumed to enforce trust boundaries.<n>We identify a recognition-enforcement gap: source-format features are linearly decodable from model activations, and models can explicitly identify forged authority when prompted.<n>We treat model self-arbitration as a capability rather than a security boundary and implement an external reference monitor combining authenticated source routing with capability-gated tool execution.
  arXiv  Detail & Related papers  (2026-08-28T16:34:05Z)
• Do GUI Agents Believe Their Eyes? Diagnosing State-Belief Reliance on Pixels versus Structure [5.717981279944186]
  We formalize visual state reliance, the attribution of a state belief to pixels, structure, or priors.<n>We measure it with paired single-channel interventions over 310 real web, mobile, and desktop probes.<n>Our central metric is the Perception-Fusion Gap, the fraction of probes a model perceives correctly yet resolves toward structure under conflict.
  arXiv  Detail & Related papers  (2026-07-05T14:33:29Z)
• From Mirage to Grounding: Towards Reliable Multimodal Circuit-to-Verilog Code Generation [16.148817642071638]
  A circuit diagram can be viewed as a visual domain-specific language for hardware.<n>Translating such diagrams into register-transfer-level(RTL) code represents an extreme reliability test for vision-to-code generation.<n>We reveal a phenomenon we call Mirage: replacing a circuit diagram with a blank image leaves Pass@k unchanged or even higher.
  arXiv  Detail & Related papers  (2026-04-30T15:01:27Z)
• 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)
• CLUE: Non-parametric Verification from Experience via Hidden-State Clustering [64.50919789875233]
  We show that correctness of a solution is encoded as a geometrically separable signature within the trajectory of hidden activations.<n>ClUE consistently outperforms LLM-as-a-judge baselines and matches or exceeds modern confidence-based methods in reranking candidates.
  arXiv  Detail & Related papers  (2025-10-02T02:14:33Z)
• WGLE:Backdoor-free and Multi-bit Black-box Watermarking for Graph Neural Networks [2.3612692427322313]
  We propose WGLE, a novel black-box watermarking paradigm for Graph Neural Networks (GNNs)<n>WGLE embeds the watermark encoding the intended information without introducing incorrect mappings that compromise the primary task.<n>Results show that WGLE achieves 100% ownership verification accuracy, an average fidelity of 0.85%, comparable against potential attacks, and low embedding overhead.
  arXiv  Detail & Related papers  (2025-06-10T09:12:00Z)
• DECIDER: Leveraging Foundation Model Priors for Improved Model Failure Detection and Explanation [18.77296551727931]
  We propose DECIDER, a novel approach that leverages priors from large language models (LLMs) and vision-language models (VLMs) to detect failures in image models. DECIDER consistently achieves state-of-the-art failure detection performance, significantly outperforming baselines in terms of the overall Matthews correlation coefficient.
  arXiv  Detail & Related papers  (2024-08-01T07:08:11Z)
• Part-Aware Transformer for Generalizable Person Re-identification [138.99827526048205]
  Domain generalization person re-identification (DG-ReID) aims to train a model on source domains and generalize well on unseen domains. We propose a pure Transformer model (termed Part-aware Transformer) for DG-ReID by designing a proxy task, named Cross-ID Similarity Learning (CSL) This proxy task allows the model to learn generic features because it only cares about the visual similarity of the parts regardless of the ID labels.
  arXiv  Detail & Related papers  (2023-08-07T06:15:51Z)
• Causal Attention for Unbiased Visual Recognition [76.87114090435618]
  Attention module does not always help deep models learn causal features that are robust in any confounding context. We propose causal attention module (CaaM) that self-annotates the confounders in unsupervised fashion. In OOD settings, deep models with CaaM outperform those without it significantly.
  arXiv  Detail & Related papers  (2021-08-19T16:45:51Z)
• Video-based Person Re-identification without Bells and Whistles [49.51670583977911]
  Video-based person re-identification (Re-ID) aims at matching the video tracklets with cropped video frames for identifying the pedestrians under different cameras. There exists severe spatial and temporal misalignment for those cropped tracklets due to the imperfect detection and tracking results generated with obsolete methods. We present a simple re-Detect and Link (DL) module which can effectively reduce those unexpected noise through applying the deep learning-based detection and tracking on the cropped tracklets.
  arXiv  Detail & Related papers  (2021-05-22T10:17:38Z)
• AvgOut: A Simple Output-Probability Measure to Eliminate Dull Responses [97.50616524350123]
  We build dialogue models that are dynamically aware of what utterances or tokens are dull without any feature-engineering. The first model, MinAvgOut, directly maximizes the diversity score through the output distributions of each batch. The second model, Label Fine-Tuning (LFT), prepends to the source sequence a label continuously scaled by the diversity score to control the diversity level. The third model, RL, adopts Reinforcement Learning and treats the diversity score as a reward signal.
  arXiv  Detail & Related papers  (2020-01-15T18:32:06Z)

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.