Not So Intrinsic: Rethinking Intrinsic Dimension Estimation in Neural Representations

NeurIPS 2025 Workshop NeurReps Submission23 Authors

Keywords: Intrinsic Dimension, Neural Representations

Abstract: The analysis of neural representation has become an integral part of research aiming to better understand the inner workings of neural networks. While there are many different approaches to investigate neural representations, an important line of research has focused on doing so through the lens of intrinsic dimensions (IDs). While this perspective has provided valuable insights and stimulated substantial follow-up research, important limitations of this approach have remained largely unaddressed. In this paper, we highlight an important discrepancy between theory and practice of IDs in neural representations, showing that common ID estimators are, in fact, not tracking the true underlying ID of the representation. We contrast this negative result with an investigation of the underlying factors that may drive commonly reported ID-related results on neural representation in the literature.

Submission Number: 23