AI News HubLIVE
原文2 min read

Ablate-to-Validate: Are Vision-Language Models Really Using Continuous Thought Tokens?

This paper proposes the Ablate-to-Validate diagnostic principle and its instantiation, the Token Replacement Test (TRT), to determine whether vision-language models (VLMs) genuinely use continuous latent tokens for reasoning. Experiments show that VLMs retain most performance gains even when token content is corrupted or replaced, indicating that accuracy improvements are a misleading proxy for latent-token reasoning.

SourcearXiv Computer VisionAuthor: Tianyi Zhang, Mahtab Bigverdi, Ranjay Krishna

[2605.21642] Ablate-to-Validate: Are Vision-Language Models Really Using Continuous Thought Tokens?

[Submitted on 20 May 2026]

Title:Ablate-to-Validate: Are Vision-Language Models Really Using Continuous Thought Tokens?

View a PDF of the paper titled Ablate-to-Validate: Are Vision-Language Models Really Using Continuous Thought Tokens?, by Tianyi Zhang and 2 other authors

View PDF HTML (experimental)

Abstract:Vision-language models (VLMs) are increasingly augmented with continuous or latent non-textual tokens intended to support "visual thinking." Despite improved task accuracy, this alone does not show that models actually use these tokens for reasoning -- gains may arise from confounds such as added context length, special-token anchoring, or training-time regularization. We formalize a diagnostic principle, Ablate-to-Validate, for testing whether latent-token content is genuinely utilized, and instantiate it as the Token Replacement Test (TRT), a standardized suite of content-replacement ablations. TRT holds the prompt, image, token budget, and decoding fixed while replacing intermediate tokens with zero, random, first-repeat, or oracle alternatives, isolating whether performance depends on token content or merely on token presence. As a controlled testbed, we study relative depth reasoning with LLaVA-13B and Qwen2.5-VL-3B, training models to predict and consume continuous or discrete depth spans across multiple frozen encoders (SigLIP2, CLIP, DINOv2) and token budgets. We additionally apply TRT to three off-the-shelf visual-thinking systems (Mirage, Mull-Tokens, CoVT) on BLINK, VSP, and CV-Bench. Across all settings, accuracy gains are a misleading proxy for latent-token reasoning: VLMs retain most improvement even when token content is corrupted or replaced, revealing a persistent gap between having a latent channel and using it as an information bottleneck. We recommend TRT as a standard diagnostic alongside accuracy for any method introducing continuous thought tokens.

Subjects:

Computer Vision and Pattern Recognition (cs.CV)

Cite as: arXiv:2605.21642 [cs.CV]

(or arXiv:2605.21642v1 [cs.CV] for this version)

https://doi.org/10.48550/arXiv.2605.21642

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Tianyi Zhang [view email] [v1] Wed, 20 May 2026 18:55:16 UTC (96 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Ablate-to-Validate: Are Vision-Language Models Really Using Continuous Thought Tokens?, by Tianyi Zhang and 2 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.CV

new | recent | 2026-05

Change to browse by:

cs

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?)