Back to AI News
arxiv

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

Researchers introduce the Multimodal Concept Lens (MC-Lens), a multimodal Jacobian lens for studying concept processing in vision-language models. The method maps intermediate vision-encoder and language-decoder states into a shared semantic space defined by the decoder's output vocabulary. Counterfactual interventions indicate that accessed concept information participates in downstream computation, and controlled VLM training shows selective access developing later than concept accessibility. Shared patterns of concept accessibility, query-dependent selectivity and cross-task intervention effects appear across VLM families and sizes despite internal differences.

AI Brief

It offers a concrete interpretability method and experimental findings for understanding how vision-language models represent and use multimodal concepts internally.

arXiv:2611.26590v1 Announce Type: new Abstract: Vision-language models (VLMs) integrate visual and textual information, but their outputs do not reveal internal concept representations. Global workspace theory (GWT) proposes that information is selected from parallel brain processes, integrated in a shared workspace, and made available to specialised processes. We introduce the Multimodal Concept Lens (MC-Lens), a multimodal Jacobian lens, to study these functions in VLM concept processing. Its vision encoder and language decoder lenses map intermediate states to a shared semantic space defined by the decoder's output vocabulary to compare concept accessibility across modalities. MC-Lens supports internal concept interventions to examine the role of accessed concepts in downstream computation. (i) Accessibility and selective access. Multimodal concept information is readable in intermediate states and selectively accessible as tasks require. We extend "readout carriers" from one name token to a token set to better capture and express concept information and support counterfactual experiments. (ii) Concepts in downstream computation. Counterfactual experiments show that accessed concept information participates in downstream computation and reasoning. (iii) Development of cross-modal alignment and selective access. We develop two J-space metrics tracking the development of cross-modal concept alignment more clearly than hidden state cosine in controlled VLM training. The experiment also shows that reliable selective access develops later than concept accessibility. (iv) Shared functions across models. Across VLM families and sizes, we find shared patterns of concept accessibility, query-dependent selectivity and cross-task intervention effects despite differences in internal organisation. These findings offer an experimental basis to study GWT in multimodal artificial systems and how it can guide cross-modal alignment improvements.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!