Representation Geometry in Language Models

Follow this topic Ask about this topic

Momentum

48 papers in the last four weeks, up 118% on the four weeks before. 0.5% of all new papers.

Jul 13Week of Sep 28

Latest papers 221

All topics

Parts-of-Speech as Emergent Categories in SAE Latent Space
Sep 24, 2026Alessandro Bondielli, Lucia Passaro, Serena Auriemma +1Sparse Autoencoders Representation Geometry in Language Models
CoLingLab, Department of Philology, Literature and Linguistics, University of Pisa · Department of Computer Science, University of Pisa
Order-Invariant Answers, Order-Sensitive Representations in Mathematical Reasoning
Sep 23, 2026Zhixu Silvia TaoMathematical Reasoning Representation Geometry in Language Models
Operations Research and Financial Engineering Princeton University Princeton, NJ 08540
Computation Over Geometry: Meaning Identity Is Computed, Not Shipped in the Embeddings
Sep 23, 2026Jiaqi DengSemantic Correspondence Cross-Encoder Reranking
Independent Researcher
The Answer-Basin Representation Hypothesis: We Are Not Probing or Steering Concepts
Sep 21, 2026Manjiang Yu, Hongji Li, Zihan Wang +5LLM Interpretability Linear Representation Hypothesis
The University of Queensland · Mohamed bin Zayed University of Artificial Intelligence · Institute of Science Tokyo
Comparing Latent Concept Formation in State Space Models and Transformers via Sparse Autoencoders
Sep 21, 2026Rithin Nagaraj, Rupa Laalasa Oruganti, Prerna Subhashchandra Kunder +1Transformer Interpretability Language Modeling
Department of Computer Science and Engineering PES University Bangalore, India
Displacement Geometry Captures Platonic Shared Reality Across Models and Modalities
Sep 21, 2026Chenming Shang, Yujin Tang, Jun Jie Ou Yang +3Cross-Modal Alignment Transfer Learning
Department of Computer Science, DARTMOUTH COLLEGE
Opinion Leader Dynamics: How Sparse Attention Shapes Token Clustering
Sep 21, 2026Jingkun Liu, Yue SongOpinion Dynamics Clustering
Tsinghua University · Shanghai Jiao Tong University
Same Outcome, Different Readout: What Does a Steerable Valence Direction in LLMs Represent?
Sep 19, 2026Weihan Li, Xinlei Chen, Yuhan Song +2Language Model Steering LLM Interpretability
The University of Tokyo · Harbin Institute of Technology, Shenzhen · The Hong Kong University of Science and Technology
Generalization through Lexical Abstraction in Transformer Models: The Case of Functional Words
Sep 17, 2026Giuseppe Samo, Vivi Nastase, Paola MerloTransformer Interpretability Word Embeddings
Idiap Research Institute, Martigny, Switzerland · University of Geneva, Swizerland
Large Language Models Develop Belief State Geometry In-Context
Sep 15, 2026Daniel Balcells, Andrew Jun Lee, Chirag Rastogi +3In-Context Learning Belief Updating in LLMs
Independent · UCLA · UIUC +2
Attention Mean Fields Predict Average Representation Dynamics and Reveal Context-Specific Computation
Sep 14, 2026Micah Adler, John W. Byers, Mark CrovellaAttention Head Analysis In-Context Learning
Department of Computer Science, Boston University · Faculty of Computing & Data Sciences, Boston University
Disentangling Representation Evolution in Transformers through Directional Decomposition
Sep 14, 2026Shwai He, Haichao Zhang, Shen YanNeural Representation Geometry Transformer Attention
University of Maryland, College Park · Northeastern University · ByteDance
Psychosis involves a deficit of information compression in connected speech
Sep 14, 2026Samuele Vallisa, Claudio Palominos, Rui He +6Computational Psycholinguistics Representation Geometry in Language Models
Grammar and Cognition Lab, Department of Translation & Language Sciences, Universitat Pompeu Fabra, Barcelona, Spain · Department of Neurosciences, Health Sciences Institute, Dokuz Eylul University, Izmir, Turkey. · Department of Psychiatry, Dokuz Eylul University, Izmir, Turkey. +4
Implicit Personality Representations in Humans and LLMs
Sep 14, 2026Yilin Geng, Omri Abend, Eduard Hovy +1Personality Modeling in Language Models Representation Geometry in Language Models
University of Melbourne · Hebrew University of Jerusalem · University of Tübingen
MAxBench: A Multinomial Concept Recovery Benchmark
Sep 14, 2026Divya Appapogu, Freya Behrens, Yonatan Belinkov +1LLM Interpretability Representation Geometry in Language Models
Department of Computer Science, Boston University, Boston, MA 02215, USA · Department of Computer Science, Technion – Israel Institute of Technology, Israel · Department of Computer Science, Kempner Institute, Harvard University, Boston, USA
Semantic Fibers and Cross-Gram Interference: A Calculus of Safety Drift in Overcomplete Representations
Sep 14, 2026Mohammed Ahnouch, Lotfi ElaachackLanguage Model Safety Evaluation LLM Safety
Université Paris 1 Paris, France · Faculty of Science and Technology of Tangier Abdelmalek Essaadi University Tangier, Morocco
How Should Reasoning Be Organized in a Transformer's Latent Space?
Sep 12, 2026Hongyu Gu, Chang Liu, Jingwen FuTransformer Attention Implicit Reasoning in Language Models
University of Science and Technology of China Hefei, China · Zhongguancun Academy Beijing, China
GEOSTEER: Geodesic Optimization for Activation Steering in Large Language Models
Sep 11, 2026Xuan Cuong Ngo, Hao Vo, Ngan LeLanguage Model Steering Representation Geometry in Language Models
University of Arkansas, Fayetteville, Arkansas, USA
K/V-Cache Interventions Dissociate Representation Alignment from Persona Expression in Decoder-Only Language Models
Sep 11, 2026Yu Sun, Mengyin Lu, Cong Feng +2KV Caching Decoder-Only Language Models
LinkedIn
The information geometry of large language models is shared, learned, and controllable
Sep 11, 2026Dario PicozziInformation Geometry Language Model Steering
Department of Physics and Astronomy, University College London (UCL), Gower Street, London WC1E 6BT, United Kingdom · London Centre for Nanotechnology, 19 Gordon St, London WC1H 0AH, United Kingdom
Global Divergence, Local Convergence: Representation Geometry in SSMs and Transformers
Sep 8, 2026Amit Ben-Artzy, Roy SchwartzTransformer Representation Geometry in Language Models
Computer Science, The Hebrew University of Jerusalem
LLM Layers Immediately Correct Each Other
Sep 7, 2026Arjun Patrawala, Jiahai Feng, Erik Jones +1Transformer Interpretability Representation Geometry in Language Models
University of California, Berkeley
A*-Thought-V2: Efficient Latent Reasoning via Geometric Dynamics of LLM
Sep 7, 2026Xiaoang Xu, Siyuan Liu, Shuo Wang +13Efficient Language Model Reasoning Continuous Latent Reasoning
Beijing University of Posts and Telecommunications · OpenBMB · The Hong Kong Polytechnic University +2
Think Wider: Mitigating Latent Rank Collapse in Implicit Chain-of-Thought Reasoning
Sep 7, 2026Yuwen Hao, Menglin YangSpectral Regularization Continuous Latent Reasoning
The Hong Kong University of Science and Technology (Guangzhou)
The Gradient Does Not See Rank: Rank-Indifference in Matrix-CODI on ProsQA
Sep 2, 2026Samuel LarsonCoT Reasoning Low-Rank Approximation
Pebble ML
The Dynamics of Continuous Mixture Collapse in Language Models
Sep 2, 2026Ali BackourContinuous Latent Reasoning Representation Geometry in Language Models
Massachusetts Institute of Technology Cambridge, MA 02139
Interpretable Symptom Vectors for Depression in a Large Language Model
Sep 1, 2026Fangyi Zhu, Ajay Subramanian, Allison Constant +3Depression Detection Representation Geometry in Language Models
Department of Psychiatry & Behavioral Sciences, Stanford University School of Medicine, Stanford, CA, USA · Wu Tsai Neurosciences Institute, Stanford University, Stanford, CA, USA · Kaiser Permanente, San Jose, CA, USA +2
Are You Thinking What I am Thinking? : Examining Conceptual Separation in Neural Architectures
Sep 1, 2026Jaee Ponde, Roshni Agarwal, Subhashis BanerjeeNeural Network Interpretability Convolutional Neural Networks
Truth Audit Labs · Karya AI · Department of Computer Science and the Centre of Digitalisation, AI, and Society, Ashoka University, Sonepat, Haryana, India
Measuring Optimal Transport in Transformer Depth
Sep 1, 2026Alexandre QuemyTransformer Neural Network Training Dynamics
Hother Labs
Late Transformer Layers Recode Syntax Canonically: Evidence from Greek Scrambling and Cross-Layer Generalisation
Aug 31, 2026Christos Nikolaos Zacharopoulos, Revekka Kyriakoglou, Chara Tsoukala +1Transformer Interpretability Language Modeling
Independent Researcher · Université Paris 8 Vincennes–Saint-Denis Paris, France · Athena Research Center, Athens, Greece +1

Sep 24, 2026, cs.CLJ/K move · Enter open · S save

Parts-of-Speech as Emergent Categories in SAE Latent Space

Alessandro Bondielli, Lucia Passaro, Serena Auriemma, Alessandro Lenci

CoLingLab, Department of Philology, Literature and Linguistics, University of Pisa · Department of Computer Science, University of Pisa

Sparse Autoencoders Representation Geometry in Language Models

Open paper AskPDF arXiv

Sparse AutoEncoders (SAEs) offer a promising way to inspect language model representations, but it is still unclear what kind of linguistic structure their latents expose. We use part-of-speech (PoS) categories as a controlled test case to study whether morpho-syntactic information is encoded by individual latents or by structured groups of features. We find that PoS distinctions are highly recoverable from SAE activations, but do not align with one-to-one latent / category mappings. This recoverability is not reducible to lexical memorisation, and Open and Closed PoS classes differ substantially. Categories are supported by compact groups of sparse latents, with substantial variation across tags. These groups remain stable on held-out data, while also showing overlap between related categories. Our results show that SAEs localise morpho-syntactic information in a distributed and category-dependent form rather than through atomic grammatical features.

Momentum

Related topics

Latest papers 221