Emergent Causal-Geometric Dynamics Across Depth in Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Haim, Shahar, McNamee, Daniel C
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916048577495040
author Haim, Shahar
McNamee, Daniel C
author_facet Haim, Shahar
McNamee, Daniel C
contents Geometric analyses of large language model (LLM) representations reveal structured variation across depth but remain fundamentally correlational with respect to token prediction formation. Meanwhile, causal interventions expose depth-dependent efficacy profiles without a unifying account of their representational dynamics. A complete account of LLM function requires explaining how representational structure evolves across depth to causally produce predictions. We synthesize these perspectives by combining geometric analysis with mechanistic interventions, explicitly centralizing depth-wise dynamics as the organizing axis for interpreting LLM function. In decoder-only LLMs, we identify a sharp transition from context-processing to prediction-forming computation, accompanied by a more gradual reorganization of representational geometry across layers. This synthesis reveals a late-layer geometric code in which angular structure parameterizes next-token distributional similarity and enables selective causal control over predictions, while representation norms encode information largely decoupled from prediction. Together, our results provide a synthesis of causal and geometric perspectives, yielding a mechanistic account of how control-relevant geometric dynamics across depth transform context into prediction in language models. This perspective reconciles previously puzzling findings and implies that layer-wise function cannot be understood or effectively intervened upon in isolation, but only within the emergent global dynamical structure of the network.
format Preprint
id arxiv_https___arxiv_org_abs_2602_04931
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Emergent Causal-Geometric Dynamics Across Depth in Large Language Models
Haim, Shahar
McNamee, Daniel C
Machine Learning
Artificial Intelligence
Geometric analyses of large language model (LLM) representations reveal structured variation across depth but remain fundamentally correlational with respect to token prediction formation. Meanwhile, causal interventions expose depth-dependent efficacy profiles without a unifying account of their representational dynamics. A complete account of LLM function requires explaining how representational structure evolves across depth to causally produce predictions. We synthesize these perspectives by combining geometric analysis with mechanistic interventions, explicitly centralizing depth-wise dynamics as the organizing axis for interpreting LLM function. In decoder-only LLMs, we identify a sharp transition from context-processing to prediction-forming computation, accompanied by a more gradual reorganization of representational geometry across layers. This synthesis reveals a late-layer geometric code in which angular structure parameterizes next-token distributional similarity and enables selective causal control over predictions, while representation norms encode information largely decoupled from prediction. Together, our results provide a synthesis of causal and geometric perspectives, yielding a mechanistic account of how control-relevant geometric dynamics across depth transform context into prediction in language models. This perspective reconciles previously puzzling findings and implies that layer-wise function cannot be understood or effectively intervened upon in isolation, but only within the emergent global dynamical structure of the network.
title Emergent Causal-Geometric Dynamics Across Depth in Large Language Models
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2602.04931