AI News HubLIVE
原文2 min read

Mirage Probes: How Vision Models Fake Visual Understanding

Vision-language models (VLMs) confidently answer image-based questions even without images, inflating benchmark scores without reflecting visual grounding. This study classifies mirages into two regimes: textual biases (relying on language priors) and spurious images (constructing false visual content in latent space). Using a contrastive probing framework and a Prior Harnessing Index (PHI), the authors show these regimes are distinct and require different mitigation strategies—text cleaning addresses only the first, while faithful grounding needs representational interventions.

SourcearXiv Computer VisionAuthor: Daniel Ben-Levi, Judah Goldfeder, Weiliang Zhao, Raz Lapid, Amit LeVi, Allen G. Roush, Ravid Shwartz-Ziv, Hod Lipson

[2606.13870] Mirage Probes: How Vision Models Fake Visual Understanding

[Submitted on 11 Jun 2026]

Title:Mirage Probes: How Vision Models Fake Visual Understanding

View a PDF of the paper titled Mirage Probes: How Vision Models Fake Visual Understanding, by Daniel Ben-Levi and 7 other authors

View PDF HTML (experimental)

Abstract:Vision-language models (VLMs) can answer image-based questions confidently, and often correctly, even when no image is provided. This mirage behavior inflates benchmark scores without reflecting visual grounding. Prior work treats this as a single failure mode. We argue it is two. Using Mirage Probes, a contrastive probing framework that pairs paraphrased question variants with matched mirage and non-mirage labels on the same image, we show that mirage behavior is linearly decodable from internal activations across residual stream, MLP, post-attention, and attention-head sites in two open-source VLMs. We demonstrate that a Naive Bayes text baseline cannot recover this signal, ruling out surface lexical confounds. Cross-benchmark separability patterns, together with a novel Prior Harnessing Index (PHI) measuring how much a model can answer from text alone, expose two distinct regimes: textual biases, where the model answers from language priors without engaging visual representations, and spurious images, where it constructs false visual content in latent space and answers as if grounded. The distinction has direct mitigation consequences: text-distribution cleaning can address the first regime but cannot reach the second, since spurious-image mirages live in the model's visual representations rather than its text. Faithful visual grounding will require interventions at the representational level.

Subjects:

Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)

Cite as: arXiv:2606.13870 [cs.CV]

(or arXiv:2606.13870v1 [cs.CV] for this version)

https://doi.org/10.48550/arXiv.2606.13870

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Daniel Ben-Levi [view email] [v1] Thu, 11 Jun 2026 19:51:44 UTC (629 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Mirage Probes: How Vision Models Fake Visual Understanding, by Daniel Ben-Levi and 7 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.CV

new | recent | 2026-06

Change to browse by:

cs cs.AI cs.LG

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)