Manifolds and Modules: How Function Develops in a Neural Foundation Model

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Bertram, Johannes, Dyballa, Luciano, Keller, T. Anderson, Kinger, Savik, Zucker, Steven W.
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908701335486464
author Bertram, Johannes
Dyballa, Luciano
Keller, T. Anderson
Kinger, Savik
Zucker, Steven W.
author_facet Bertram, Johannes
Dyballa, Luciano
Keller, T. Anderson
Kinger, Savik
Zucker, Steven W.
contents Foundation models have shown remarkable success in fitting biological visual systems; however, their black-box nature inherently limits their utility for understanding brain function. Here, we peek inside a SOTA foundation model of neural activity (Wang et al., 2025) as a physiologist might, characterizing each 'neuron' based on its temporal response properties to parametric stimuli. We analyze how different stimuli are represented in neural activity space by building decoding manifolds, and we analyze how different neurons are represented in stimulus-response space by building neural encoding manifolds. We find that the different processing stages of the model (i.e., the feedforward encoder, recurrent, and readout modules) each exhibit qualitatively different representational structures in these manifolds. The recurrent module shows a jump in capabilities over the encoder module by 'pushing apart' the representations of different temporal stimulus patterns; while the readout module achieves biological fidelity by using numerous specialized feature maps rather than biologically plausible mechanisms. Overall, we present this work as a study of the inner workings of a prominent neural foundation model, gaining insights into the biological relevance of its internals through the novel analysis of its neurons' joint temporal response patterns.
format Preprint
id arxiv_https___arxiv_org_abs_2512_07869
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Manifolds and Modules: How Function Develops in a Neural Foundation Model
Bertram, Johannes
Dyballa, Luciano
Keller, T. Anderson
Kinger, Savik
Zucker, Steven W.
Neurons and Cognition
Artificial Intelligence
Foundation models have shown remarkable success in fitting biological visual systems; however, their black-box nature inherently limits their utility for understanding brain function. Here, we peek inside a SOTA foundation model of neural activity (Wang et al., 2025) as a physiologist might, characterizing each 'neuron' based on its temporal response properties to parametric stimuli. We analyze how different stimuli are represented in neural activity space by building decoding manifolds, and we analyze how different neurons are represented in stimulus-response space by building neural encoding manifolds. We find that the different processing stages of the model (i.e., the feedforward encoder, recurrent, and readout modules) each exhibit qualitatively different representational structures in these manifolds. The recurrent module shows a jump in capabilities over the encoder module by 'pushing apart' the representations of different temporal stimulus patterns; while the readout module achieves biological fidelity by using numerous specialized feature maps rather than biologically plausible mechanisms. Overall, we present this work as a study of the inner workings of a prominent neural foundation model, gaining insights into the biological relevance of its internals through the novel analysis of its neurons' joint temporal response patterns.
title Manifolds and Modules: How Function Develops in a Neural Foundation Model
topic Neurons and Cognition
Artificial Intelligence
url https://arxiv.org/abs/2512.07869