The Changing Geometry of Grammar: Dimensionality and Neighborhood Reorganization across Transformer Layers

arXiv cs.CL Papers

Summary

This paper investigates how the grammatical role of tokens shapes the geometry of transformer representations across layers, finding distinct evolution patterns in encoder versus decoder models.

arXiv:2608.25166v1 Announce Type: new Abstract: Transformer representations describe trajectories through high-dimensional vector spaces, which are shaped dynamically as tokens incorporate relational context across layers. Such data tend to concentrate on lower-dimensional sub-manifolds, a form of compression quantified by the Intrinsic Dimensionality (ID), the minimum number of independent variables needed to represent them without significant information loss. In this work, we ask whether the grammatical role of tokens, as marked by their part-of-speech (PoS) tag, shapes the local geometry of this manifold. To this end: (1) We investigate the layer-wise evolution of ID, finding that closed-class items expand earlier and collapse sooner than open-class ones; (2) We show its expansion and contraction to be explained by changes in the neighborhood structure, and hence in the relations between words within a sentence; (3) We compare encoders (ModernBERT, bigbird-roberta-large) and decoders (gemma-2-2B, Llama-3.2-3B), finding that the two families evolve differently across layers, consistently with how each integrates context;(4) We show that geometric features alone recover a token's grammatical role, and use them to interpret how the semantic content of each PoS evolves across layers in a downstream classification task.
Original Article
View Cached Full Text

Cached at: 08/27/26, 09:16 AM

# The Changing Geometry of Grammar: Dimensionality and Neighborhood Reorganization across Transformer Layers
Source: [https://arxiv.org/abs/2608.25166](https://arxiv.org/abs/2608.25166)
[View PDF](https://arxiv.org/pdf/2608.25166)

> Abstract:Transformer representations describe trajectories through high\-dimensional vector spaces, which are shaped dynamically as tokens incorporate relational context across layers\. Such data tend to concentrate on lower\-dimensional sub\-manifolds, a form of compression quantified by the Intrinsic Dimensionality \(ID\), the minimum number of independent variables needed to represent them without significant information loss\. In this work, we ask whether the grammatical role of tokens, as marked by their part\-of\-speech \(PoS\) tag, shapes the local geometry of this manifold\. To this end: \(1\) We investigate the layer\-wise evolution of ID, finding that closed\-class items expand earlier and collapse sooner than open\-class ones; \(2\) We show its expansion and contraction to be explained by changes in the neighborhood structure, and hence in the relations between words within a sentence; \(3\) We compare encoders \(ModernBERT, bigbird\-roberta\-large\) and decoders \(gemma\-2\-2B, Llama\-3\.2\-3B\), finding that the two families evolve differently across layers, consistently with how each integrates context;\(4\) We show that geometric features alone recover a token's grammatical role, and use them to interpret how the semantic content of each PoS evolves across layers in a downstream classification task\.

## Submission history

From: Federico Ravenda \[[view email](https://arxiv.org/show-email/b90e13cf/2608.25166)\] **\[v1\]**Tue, 25 Aug 2026 21:27:21 UTC \(4,586 KB\)

Similar Articles

Geometric and Behavioral Stratification in Transformer Residual Streams

arXiv cs.LG

This paper investigates the residual stream geometry of trained transformers, showing that the prediction direction acts as a privileged anchor that stratifies residual-stream variation into narrow, readout-relevant and broad, computational regions across 18 models. The findings have implications for interpretability and evaluation.

An Analysis of Residual-Stream Geometry Across Transformer Depth

arXiv cs.LG

This paper proposes a geometric analysis of transformer residual streams across depth, using relative displacement and orthogonal Procrustes analysis to reveal structured regularities in six instruction-tuned models on code generation and translation tasks.