mlsanityData to 14 Oct 2026

Layers at Similar Depths Generate Similar Activations Across LLM Architectures

arXiv 2504.08775 ↗First posted 3 Apr 2025cs.CL cs.AIcited by 12

Not on the map. Why?

Abstract (arXiv)

How do the latent spaces used by independently-trained LLMs relate to one another? We study the nearest neighbor relationships induced by activations at different layers of 24 open-weight LLMs, and find that they 1) tend to vary from layer to layer within a model, and 2) are approximately shared between corresponding layers of different models. Claim 2 shows that these nearest neighbor relationships are not arbitrary, as they are shared across models, but Claim 1 shows that they are not "obvious" either, as there is no single set of nearest neighbor relationships that is universally shared. Together, these suggest that LLMs generate a progression of activation geometries from layer to layer, but that this entire progression is largely shared between models, stretched and squeezed to fit into different architectures.

Key facts. Cited by 12 papers in the map (data through 14 Oct 2026). Builds on 23 of its 36 references that are in the map. First posted on arXiv 3 Apr 2025.

Builds on

23 of 36 references.

References · 23

  1. 2017Attention Is All You Needcited by 60,361
  2. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  3. 2024Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phonecited by 1,917
  4. 2019Similarity of Neural Network Representations Revisitedcited by 1,755
  5. 2023Instruction-Following Evaluation for Large Language Modelscited by 1,228
  6. 2019BERT Rediscovers the Classical NLP Pipelinecited by 1,059
  7. 2023Instruction Tuning for Large Language Models: A Surveycited by 623
  8. 2020Do Wide and Deep Networks Learn the Same Things? Uncovering How Neural Network Representations Vary with Width and Depthcited by 245
  9. 2020Probing Pretrained Language Models for Lexical Semanticscited by 156
  10. 2024The Remarkable Robustness of LLMs: Stages of Inference?cited by 146
  11. 2023Similarity of Neural Network Models: A Survey of Functional and Representational Measurescited by 131
  12. 2024A Practical Review of Mechanistic Interpretability for Transformer-Based Language Modelscited by 118
11 more
  1. 2024Universal Neurons in GPT2 Language Modelscited by 102
  2. 2020Analyzing Individual Neurons in Pre-trained Language Modelscited by 97
  3. 2024Emergence of a High-Dimensional Abstraction Phase in Language Transformerscited by 56
  4. 2020Similarity Analysis of Contextual Word Representation Modelscited by 54
  5. 2024From Tokens to Words: On the Inner Lexicon of LLMscited by 48
  6. 2019Blackbox meets blackbox: Representational Similarity and Stability Analysis of Neural Language Models and Brainscited by 38
  7. 2024Towards Universality: Studying Mechanistic Similarity Across Language Model Architecturescited by 31
  8. 2023Understanding the Inner Workings of Language Models Through Representation Dissimilaritycited by 12
  9. 2022On the Origins of the Block Structure Phenomenon in Neural Network Representationscited by 10
  10. 2023Towards Measuring Representational Similarity of Large Language Modelscited by 6
  11. 2024Investigating Representation Universality: Case Study on Genealogical Representationscited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 6; 2026: 6.

12 papers.

  1. 2025Harnessing the Universal Geometry of Embeddingscited by 55
  2. 2025Correlated Errors in Large Language Modelscited by 45
  3. 2026Differential syntactic and semantic encoding in LLMscited by 7
  4. 2025Signal in the Noise: Polysemantic Interference Transfers and Predicts Cross-Model Influencecited by 4
  5. 2025Dynamic Weight Grafting: Localizing Finetuned Factual Knowledge in Transformerscited by 3
  6. 2025Large Language Models Show Signs of Alignment with Human Neurocognition During Abstract Reasoningcited by 2
  7. 2025Mapping Overlaps in Benchmarks through Perplexity in the Wildcited by 2
  8. 2026Characterizing Human Semantic Navigation in Concept Production as Trajectories in Embedding Spacecited by 2
  9. 2026Contravariance Theory: Strong Alignment for Minimal Solutions to Hard Taskscited by 2
  10. 2026Gradient Smoothing: Coupling Layer-wise Updates for Improved Optimizationcited by 0
  11. 2026SAEVerbalizer: Generating Explanations for Sparse Autoencoder Features via Representation Verbalizationcited by 0
  12. 2026Global Divergence, Local Convergence: Representation Geometry in SSMs and Transformerscited by 0