Probing Visual Concepts in Lightweight Vision-Language Models for Automated Driving

• URL: http://arxiv.org/abs/2603.06054v1
• Date: Fri, 06 Mar 2026 09:07:57 GMT
• Title: Probing Visual Concepts in Lightweight Vision-Language Models for Automated Driving
• Abstract summary: Vision-Language Models (VLMs) are used in automated driving applications.<n>These models often fail on simple visual questions that are highly relevant to automated driving.<n>We show that concepts such as the presence of an object or agent in a scene are explicitly and linearly encoded.<n>Other spatial visual concepts, such as the orientation of an object or agent, are only implicitly encoded by the spatial structure retained by the vision encoder.<n>Finally, we show that increasing the distance of the object in question quickly degrades the linear separability of the corresponding visual concept.
• Score: 3.333320380836246
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The use of Vision-Language Models (VLMs) in automated driving applications is becoming increasingly common, with the aim of leveraging their reasoning and generalisation capabilities to handle long tail scenarios. However, these models often fail on simple visual questions that are highly relevant to automated driving, and the reasons behind these failures remain poorly understood. In this work, we examine the intermediate activations of VLMs and assess the extent to which specific visual concepts are linearly encoded, with the goal of identifying bottlenecks in the flow of visual information. Specifically, we create counterfactual image sets that differ only in a targeted visual concept and then train linear probes to distinguish between them using the activations of four state-of-the-art (SOTA) VLMs. Our results show that concepts such as the presence of an object or agent in a scene are explicitly and linearly encoded, whereas other spatial visual concepts, such as the orientation of an object or agent, are only implicitly encoded by the spatial structure retained by the vision encoder. In parallel, we observe that in certain cases, even when a concept is linearly encoded in the model's activations, the model still fails to answer correctly. This leads us to identify two failure modes. The first is perceptual failure, where the visual information required to answer a question is not linearly encoded in the model's activations. The second is cognitive failure, where the visual information is present but the model fails to align it correctly with language semantics. Finally, we show that increasing the distance of the object in question quickly degrades the linear separability of the corresponding visual concept. Overall, our findings improve our understanding of failure cases in VLMs on simple visual tasks that are highly relevant to automated driving.

Related papers

• Seeing or Knowing? Visual Context Sensitivity in Multimodal Large Language Models [14.5102059776721]
  Multimodal Large Language Models (MLLMs) achieve strong performance by integrating visual inputs with the rich priors of pretrained language models.<n>However, they often fail on vision-centric tasks, especially when visual evidence conflicts with pretrained knowledge.<n>We explore these failures using two diagnostic paradigms: (1) probing whether visual information is available, via image reconstruction, and (2) measuring multimodal context sensitivity.
  arXiv  Detail & Related papers  (2026-07-28T22:52:34Z)
• Diagnosing Visual Ignorance in Vision-Language Models [29.2851901986069]
  Vision-Language Models (VLMs) frequently rely on language priors, producing confident answers that are weakly grounded in visual evidence.<n>In this work, we study language-prior reliance from both mechanistic and behavioral perspectives.<n>We find that language-prior reliance is a systematic routing failure affecting both model internals and benchmark validity.
  arXiv  Detail & Related papers  (2026-06-05T04:16:01Z)
• AwareVLN: Reasoning with Self-awareness for Vision-Language Navigation [78.80536515102305]
  Vision-and-Language Navigation (VLN) requires an agent to ground language instructions to its own movement within a visual environment.<n>We propose AwareVLN, a novel framework that equips the navigation model with a self-aware reasoning mechanism.<n>Our approach features two key innovations: (1) a structural reasoning module that fosters spatial and task-oriented self-awareness, and (2) an automatic data engine with progress division for effective training.
  arXiv  Detail & Related papers  (2026-05-21T17:58:26Z)
• EgoDyn-Bench: Evaluating Ego-Motion Understanding in Vision-Centric Foundation Models for Autonomous Driving [40.67852883277799]
  Vision-Language Models (VLMs) have advanced highlevel reasoning in autonomous driving.<n>But their ability to ground this reasoning in the underlying physics of ego-motion remains poorly understood.<n>We introduce EgoDyn-Bench, a diagnostic benchmark for evaluating the semantic ego-motion understanding of vision-centric foundation models.
  arXiv  Detail & Related papers  (2026-04-22T07:49:02Z)
• Through the Lens of Contrast: Self-Improving Visual Reasoning in VLMs [60.93949629734977]
  We propose Visual Contrastive Self-Taught Reasoner (VC-STaR) to mitigate hallucinations in model-generated rationales.<n>We collect a diverse suite of VQA datasets, curate contrastive pairs according to multi-modal similarity, and generate rationales using VC-STaR.<n>Extensive experiments show that VC-STaR not only outperforms existing self-improving approaches but also surpasses models finetuned on the SoTA visual reasoning datasets.
  arXiv  Detail & Related papers  (2026-03-03T03:18:31Z)
• The Geometry of Representational Failures in Vision Language Models [5.7337123720860435]
  Vision-Language Models (VLMs) exhibit puzzling failures in multi-object visual tasks.<n>These errors mirror human cognitive constraints, such as the "Binding Problem"<n>We propose a mechanistic insight by analyzing the representational geometry of open-weight VLMs.
  arXiv  Detail & Related papers  (2026-02-02T12:20:04Z)
• Linear Mechanisms for Spatiotemporal Reasoning in Vision Language Models [16.73630874846666]
  Visual/geometrical and textual representations of spatial structure must be combined at some point in Vision Language Models (VLMs)<n>We show thatVLMs encode object locations by linearly binding itspatialtemporal IDs to textual activations, then perform reasoning via language tokens.<n>We extend our analysis to video VLMs and identify an analogous linear temporal ID mechanism.
  arXiv  Detail & Related papers  (2026-01-18T23:48:38Z)
• Vision-Language-Action Models for Autonomous Driving: Past, Present, and Future [125.92052530850425]
  Vision-Language-Action (VLA) frameworks integrate perception with language-grounded decision making.<n>VLA frameworks offer a pathway toward more interpretable, generalizable, and human-aligned driving policies.<n>This work aims to establish a coherent foundation for advancing human-compatible autonomous driving systems.
  arXiv  Detail & Related papers  (2025-12-18T16:57:44Z)
• SAVE: Sparse Autoencoder-Driven Visual Information Enhancement for Mitigating Object Hallucination [48.601385640941935]
  We propose SAVE, a framework that mitigates hallucination by steering the model along Sparse Autoencoder latent features.<n>A binary object-presence question-answering probe identifies the SAE features most indicative of the model's visual information processing.<n>With its simple design, SAVE outperforms state-of-the-art training-free methods on standard benchmarks.
  arXiv  Detail & Related papers  (2025-12-08T17:20:07Z)
• Self-Rewarding Vision-Language Model via Reasoning Decomposition [49.784411666601905]
  Vision-Language Models (VLMs) often suffer from visual hallucinations, saying things that are not actually in the image, and language shortcuts.<n>We introduce Vision-SR1, a self-rewarding method that improves visual reasoning without relying on external visual supervisions.<n>Our experiments demonstrate that Vision-SR1 improves visual reasoning, mitigates visual hallucinations, and reduces reliance on language shortcuts.
  arXiv  Detail & Related papers  (2025-08-27T08:01:03Z)
• FastDriveVLA: Efficient End-to-End Driving via Plug-and-Play Reconstruction-based Token Pruning [75.80110543049783]
  We propose FastDriveVLA, a reconstruction-based vision token pruning framework for autonomous driving.<n>A novel foreground adversarial-background reconstruction strategy is designed to train ReconPruner for the visual encoder of VLA models.<n>Our approach achieves state-of-the-art results on the nuScenes open-loop planning benchmark across different pruning ratios.
  arXiv  Detail & Related papers  (2025-07-31T07:55:56Z)
• ViCrit: A Verifiable Reinforcement Learning Proxy Task for Visual Perception in VLMs [98.27348724529257]
  We introduce ViCrit (Visual Caption Hallucination Critic), an RL proxy task that trains VLMs to localize a subtle, synthetic visual hallucination injected into paragraphs of human-written image captions.<n>Models trained with the ViCrit Task exhibit substantial gains across a variety of vision-language models benchmarks.
  arXiv  Detail & Related papers  (2025-06-11T19:16:54Z)

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.