# Emergent effects of scaling on the functional hierarchies within large language models

arXiv: https://arxiv.org/abs/2501.07359 · HTML: https://mlsanity.com/paper/2501.07359/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 3 of its 4 references that are in the map.
- First posted on arXiv 13 Jan 2025.

## Builds on (most cited first, 3 of 3)

- [Emergence of a High-Dimensional Abstraction Phase in Language Transformers](https://mlsanity.com/paper/2405.15471/) (2024)
- [Fantastic Semantics and Where to Find Them: Investigating Which Layers of Generative LLMs Reflect Lexical Semantics](https://mlsanity.com/paper/2403.01509/) (2024)
- [Investigating semantic subspaces of Transformer sentence embeddings through linear structural probing](https://mlsanity.com/paper/2310.11923/) (2023)

## Built on by (most cited first, 1 of 1)

- [Visualising Information Flow in Word Embeddings with Diffusion Tensor Imaging](https://mlsanity.com/paper/2601.05713/) (2026)

## Abstract (arXiv)

Large language model (LLM) architectures are often described as functionally hierarchical: Early layers process syntax, middle layers begin to parse semantics, and late layers integrate information. The present work revisits these ideas. This research submits simple texts to an LLM (e.g., "A church and organ") and extracts the resulting activations. Then, for each layer, support vector machines and ridge regressions are fit to predict a text's label and thus examine whether a given layer encodes some information. Analyses using a small model (Llama-3.2-3b; 28 layers) partly bolster the common hierarchical perspective: Item-level semantics are most strongly represented early (layers 2-7), then two-item relations (layers 8-12), and then four-item analogies (layers 10-15). Afterward, the representation of items and simple relations gradually decreases in deeper layers that focus on more global information. However, several findings run counter to a steady hierarchy view: First, although deep layers can represent document-wide abstractions, deep layers also compress information from early portions of the context window without meaningful abstraction. Second, when examining a larger model (Llama-3.3-70b-Instruct), stark fluctuations in abstraction level appear: As depth increases, two-item relations and four-item analogies initially increase in their representation, then markedly decrease, and afterward increase again momentarily. This peculiar pattern consistently emerges across several experiments. Third, another emergent effect of scaling is coordination between the attention mechanisms of adjacent layers. Across multiple experiments using the larger model, adjacent layers fluctuate between what information they each specialize in representing. In sum, an abstraction hierarchy often manifests across layers, but large models also deviate from this structure in curious ways.