The Platonic Representation Hypothesis

# The Platonic Representation Hypothesis

**arXiv ID:** 2405.07987
**Authors:** Minyoung Huh, Brian Cheung, Tongzhou Wang, Phillip Isola
**Published:** 2024-05-13T17:58:30Z
**Abstract:**
We argue that representations in AI models, particularly deep networks, are converging. First, we survey many examples of convergence in the literature: over time and across multiple domains, the ways by which different neural networks represent data are becoming more aligned. Next, we demonstrate convergence across data modalities: as vision models and language models get larger, they measure distance between datapoints in a more and more alike way. We hypothesize that this convergence is driving toward a shared statistical model of reality, akin to Plato's concept of an ideal reality. We term such a representation the platonic representation and discuss several possible selective pressures toward it. Finally, we discuss the implications of these trends, their limitations, and counterexamples to our analysis.

## Skill Description

This skill is generated from the arXiv paper: The Platonic Representation Hypothesis (2405.07987).

## How to Use

[To be filled in by the user or by future automation]

## References

- [arXiv:2405.07987](http://arxiv.org/abs/2405.07987v5)