The Geometry of Categorical and Hierarchical Concepts in Large Language Models
| Properties | |
|---|---|
| authors | Kiho Park, Yo Joong Choe, Yibo Jiang, Victor Veitch |
| year | 2025 |
| url | http://arxiv.org/abs/2406.01506 |
Abstract
The linear representation hypothesis is the informal idea that semantic concepts are encoded as linear directions in the representation spaces of large language models (LLMs). Previous work has shown how to make this notion precise for representing binary concepts that have natural contrasts (e.g., {male, female}) as directions in representation space. However, many natural concepts do not have natural contrasts (e.g., whether the output is about an animal). In this work, we show how to extend the formalization of the linear representation hypothesis to represent features (e.g., is_animal) as vectors. This allows us to immediately formalize the representation of categorical concepts as polytopes in the representation space. Further, we use the formalization to prove a relationship between the hierarchical structure of concepts and the geometry of their representations. We validate these theoretical results on the Gemma and LLaMA-3 large language models, estimating representations for 900+ hierarchically related concepts using data from WordNet.
Notes¶
We show how to move from representations of binary concepts as directions to representa-tions as vectors.
Figure 5 shows the cosine similarity ¯ℓparent and¯ℓchild − ¯ℓparent for the WordNet features. As predicted, this value is close to 0