Open Brain
arXivPaperNeeds Review

Probing the "Conscious" Mind: Selective Access and Global Workspace in Vision-Language Models

A study of vision-language models finds that concepts can be present in internal representations before the model can reliably select them for a task. Its analysis tool compares concept accessibility across image and text pathways.

arXiv · Nov 19, 2026
Open Source Open MarkdownOpen JSON
Source Summary

The paper introduces **MC-Lens**, which maps vision-encoder and language-decoder states into a shared semantic space and supports interventions on accessed concepts. Experiments find that concept information is readable internally, while reliable selective access develops later; accessed concepts also affect downstream reasoning.

Practical Implication

For builders using vision-language models, output quality may depend on whether the model can select relevant visual concepts for a particular task, not only whether those concepts are represented. MC-Lens offers a research method for probing that distinction.

Agent-Ready Context
The paper introduces **MC-Lens**, which maps vision-encoder and language-decoder states into a shared semantic space and supports interventions on accessed concepts. Experiments find that concept information is readable internally, while reliable selective access develops later; accessed concepts also affect downstream reasoning.

For builders using vision-language models, output quality may depend on whether the model can select relevant visual concepts for a particular task, not only whether those concepts are represented. MC-Lens offers a research method for probing that distinction.

The findings describe controlled experiments across model families and sizes. The material does not establish that these measurements predict performance in deployed applications.
Connected Context · Feed7 Judgment

MC-Lens separates whether a vision-language model contains readable information about a concept from whether it can selectively access that concept for a task. The interventions also connect accessed concepts to downstream reasoning. This offers a way to probe a possible source of task-specific failure, but the controlled findings do not establish deployed performance.

Context Map
modelimageresearch#reasoning
Uncertainty
The findings describe controlled experiments across model families and sizes. The material does not establish that these measurements predict performance in deployed applications.