mlsanityData to 7 Oct 2026

Differential syntactic and semantic encoding in LLMs

arXiv 2601.04765 ↗First posted 8 Jan 2026cs.CL cs.AI cs.LGcited by 6

Not on the map. Why?

Abstract (arXiv)

We study how syntactic and semantic information is encoded in inner layer representations of Large Language Models (LLMs), focusing on the very large DeepSeek-V3. We find that, by averaging hidden-representation vectors of sentences sharing syntactic structure or meaning, we obtain vectors that capture a significant proportion of the syntactic and semantic information contained in the representations. In particular, subtracting these syntactic and semantic ``centroids'' from sentence vectors strongly affects their similarity with syntactically and semantically matched sentences, respectively, suggesting that syntax and semantics are, at least partially, linearly encoded. We also find that the cross-layer encoding profiles of syntax and semantics are different, and that the two signals can to some extent be decoupled, suggesting differential encoding of these two types of linguistic information in LLM representations.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Builds on 40 of its 78 references that are in the map. First posted on arXiv 8 Jan 2026.

Builds on

40 of 78 references.

References · 40

  1. 2024Qwen2 Technical Reportcited by 2,076
  2. 2023Pythia: A Suite for Analyzing Large Language Models Across Training and Scalingcited by 1,734
  3. 2019Similarity of Neural Network Representations Revisitedcited by 1,695
  4. 2025Gemma 3 Technical Reportcited by 1,686
  5. 2023SGLang: Efficient Execution of Structured Language Model Programscited by 1,107
  6. 2019BERT Rediscovers the Classical NLP Pipelinecited by 1,046
  7. 2023The Linear Representation Hypothesis and the Geometry of Large Language Modelscited by 679
  8. 2019Designing and Interpreting Probes with Control Taskscited by 608
  9. 2019Linguistic Knowledge and Transferability of Contextual Representationscited by 382
  10. 2023Explainability for Large Language Models: A Surveycited by 331
  11. 2025Layer by Layer: Uncovering Hidden Representations in Language Modelscited by 293
  12. 2024Massive Activations in Large Language Modelscited by 277
28 more
  1. 2017All-but-the-Top: Simple and Effective Postprocessing for Word Representationscited by 205
  2. 2021Revisiting Model Stitching to Compare Neural Representationscited by 203
  3. 2023Linearity of Relation Decoding in Transformer Language Modelscited by 190
  4. 2024The Geometry of Categorical and Hierarchical Concepts in Large Language Modelscited by 151
  5. 2023Quantizable Transformers: Removing Outliers by Helping Attention Heads Do Nothingcited by 147
  6. 2023Getting aligned on representational alignmentcited by 142
  7. 2023Similarity of Neural Network Models: A Survey of Functional and Representational Measurescited by 128
  8. 2021BERT Busters: Outlier Dimensions that Disrupt Transformerscited by 110
  9. 2024A Primer on the Inner Workings of Transformer-based Language Modelscited by 83
  10. 2023Improving Activation Steering in Language Models with Mean-Centringcited by 64
  11. 2025Large Language Models Share Representations of Latent Grammatical Concepts Across Typologically Diverse Languagescited by 57
  12. 2024Emergence of a High-Dimensional Abstraction Phase in Language Transformerscited by 56
  13. 2025How Linguistics Learned to Stop Worrying and Love the Language Modelscited by 50
  14. 2024Characterizing Truthfulness in Large Language Model Generations with Local Intrinsic Dimensioncited by 45
  15. 2020Language Through a Prism: A Spectral Approach for Multiscale Language Representationscited by 39
  16. 2021Disentangling Syntax and Semantics in the Brain with Deep Networkscited by 39
  17. 2025Shared Global and Local Geometry of Language Model Embeddingscited by 34
  18. 2024Decoding Probing: Revealing Internal Linguistic Structures in Neural Language Models using Minimal Pairscited by 24
  19. 2024A polar coordinate system represents syntax in large language modelscited by 24
  20. 2021Ranking the information content of distance measurescited by 22
  21. 2024Evidence from fMRI Supports a Two-Phase Abstraction Process in Language Modelscited by 13
  22. 2024Concept Space Alignment in Multilingual LLMscited by 13
  23. 2025Layers at Similar Depths Generate Similar Activations Across LLM Architecturescited by 11
  24. 2025On the Emergence of Linear Analogies in Word Embeddingscited by 9
  25. 2025A quantitative analysis of semantic information in deep representations of text and imagescited by 7
  26. 2025Learning the Wrong Lessons: Syntactic-Domain Spurious Correlations in Language Modelscited by 6
  27. 2024Automatic feature selection and weighting in molecular systems using Differentiable Information Imbalancecited by 4
  28. 2025Probing Syntax in Large Language Models: Successes and Remaining Challengescited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 6.

6 papers.

  1. 2026Deriving Neural Scaling Laws from the statistics of natural languagecited by 18
  2. 2026The Myth of Expert Specialization in MoEs: Why Routing Reflects Geometry, Not Necessarily Domain Expertisecited by 9
  3. 2026The Grammar of Transformers: A Systematic Review of Interpretability Research on Syntactic Knowledge in Language Modelscited by 1
  4. 2026From Lemmas to Dependencies: What Signals Drive Light Verbs Classification?cited by 1
  5. 2026The ACUTE Protocol: Operationalizing Language Model Activations for Better Calibration, Utility, and Trustcited by 1
  6. 2026Computation Over Geometry: Meaning Identity Is Computed, Not Shipped in the Embeddingscited by 0