Representation Probing

Ask about this topic

Overview

Representation probing studies what information is encoded in a model’s internal activations and how that information changes across layers, tasks, and training. Common approaches include linear and nonlinear classifiers, feature reconstruction, sparse-neuron and geometric analyses, and interventions that test whether decoded signals affect behavior. Probes are used across language, vision, speech, biology, robotics, and other domains to assess learned features, monitor reliability, and compare models. A central open tension is that information can be easy to decode yet fail to guide a model’s outputs, so strong claims require controls for input-based shortcuts, distribution shift, and causal reliance.

Latest papers

CardsList

56 papers11 papers in the last four weeks, up 57% on the four weeks before. 0.1% of all new papers.

  1. Structured Representation Probing for Risk Anticipation in End-to-End Autonomous Driving

    Oct 27, 2026Jiawei Liang, Tianrui Lou, Le Yang +3Autonomous DrivingAutonomous Driving Safety Evaluation

    School of Cyber Science and Technology, Shenzhen Campus of Sun Yat-sen University, Shenzhen 518107, China · School of Mathematics, Sun Yat-sen University, Guangzhou 510275, China

  2. Probing the phon* representations of the Mimi codebooks

    Oct 26, 2026Artem Saloev, Nicolas Ballier, Erin PacquetetNeural Audio CodecsDiscrete Speech Representation Learning

    ALTAE, Université Paris Cité, F-75013 Paris, France · SCIAM, 10 rue de Penthièvre, F-75008 Paris, France

  3. ProMo3D: Probing Motion Cues in Frozen 3D Foundation Models

    Oct 24, 2026Xiaoang Zhang, Mert Kiray, Yordanka Velikova +2Motion Segmentation3D Foundation Models

    Technical University of Munich (TUM), Germany · Munich Center for Machine Learning (MCML), Germany · Obsphera, Munich, Germany

  4. Tropical Cyclone Attributes in Sparse Autoencoder Features of an AI Weather Model

    Oct 20, 2026Jiapei Liu, Ruidong Li, Ting Sun +1Representation ProbingWeather Forecasting

    State Key Laboratory of Hydro-science and Engineering, Department of Hydraulic Engineering, Tsinghua University, Beijing 100084, China · Department of Risk and Disaster Reduction, University College London, London WC1E 6BT, UK

  5. Taking the PULSE of Medical Reasoning: Step-Level Retrieval Timing for Knowledge-Grounded Inference

    Oct 14, 2026Danyang Zhao, Sendong Zhao, Yanrui Du +6Adaptive RetrievalEvidence Retrieval

    Harbin Institute of Technology

  6. DANGERLENS: Probing Dangerous Representations Hidden by Safety Alignment

    Oct 14, 2026Tianzhuo Yang, Shaoheng Yan, Lin Lu +5LLM SafetyLLM Safety Alignment

    Institute for Artificial Intelligence, Peking University · JIUTIAN Research, China Mobile

  7. How High Is 0.6? Floors, Ceilings, and Headroom in Interpretability Probing

    Oct 6, 2026Pranjal GargRepresentation ProbingLanguage Model Probing

    Independent Researcher

  8. Learning to Read the Contextual Tokens in Diffusion Transformers

    Oct 5, 2026Omer Dahary, Etai Sella, Hadar Averbuch-Elor +2Diffusion Model AlignmentDiffusion Transformer

    Tel Aviv University · Cornell University

  9. Readout Blindness: VLM Scores Miss the Spatial Direction Their Frozen Encoders Retain

    Oct 5, 2026Guangyuan Li, Tianming Du, Yan Jiang +2VLM EvaluationVision-Language Grounding

    ELLIS Institute Finland · Aalto University · University of Oulu +1

  10. Loss-Invariant Projections as Passive Probes of Learned Representations

    Oct 5, 2026Akshay Chandrasekhar, Pavlo MelnykNeural Representation GeometryVisual Representation Learning

    Linköping University

  11. What Does an Observability Foundation Model Know?

    Oct 4, 2026Dhyey Dharmendrakumar Mavani, Rian Atri, Tairan JiTime Series ForecastingRepresentation Probing

    Independent · Keiji AI

  12. Ghost in the Encoder: Decodable Artist Identity Representations in Lyrics-to-Song Generation

    Sep 30, 2026Arhan Vohra, Choenden Kyirong, Laura Ibáñez-Martínez +1Text-to-Music GenerationAudio Diffusion Models

    Music Technology Group, Universitat Pompeu Fabra

  13. Dynamical Parameters: An Interpretability Framework for Time-Series Foundation Models

    Sep 28, 2026Kang Yang, Gaofeng Dong, Liying Han +1Explainable Time Series ForecastingRepresentation Probing

    Department of Electrical and Computer Engineering University of California, Los Angeles

  14. Decodable In-Context State and Model Output Across Training

    Sep 25, 2026Manas Venkata Sai Ravulapalli, Samrath Singh ChadhaLLM InterpretabilityActivation Steering

    Efficient Computation Inc.

  15. Hallucination Neurons and Where to Find Them: An Investigation into the existence of Hallucination Neurons

    Sep 24, 2026Huseyin Cavus, Sebin Sabu, Joshua Spear +2LLM InterpretabilityHallucination in Language Models

    Trakya University, Edirne, Türkiye · DRIVE, Great Ormond Street Hospital for Children NHS Foundation Trust, London, UK

  16. TRIPROBE: Probing Task Separability Beyond Classification for XAI

    Sep 16, 2026Amirhossein Sadough, Freek Hens, Aleksa Bokšan +2Multi-Task LearningRepresentation Probing

    Machine Learning and Neural Computing Radboud University Nijmegen, Netherlands · Delft University of Technology Delft, Netherlands · Unconventional Computing Lab University of the West of England (UWE) Bristol, United Kingdom

  17. FARM: Reading Failure Signals from the Internal Predictive States of a Frozen Robotic World Model

    Sep 10, 2026Haoran Pei, Mingrui Luo, Senbao Wang +4Robot Failure DetectionRobotics

    Institute of Automation, Chinese Academy of Sciences, Beijing, China · Harbin Institute of Technology, Harbin, China

  18. Do Agents Know When They Succeed? Calibrating Agent Confidence from Internal Representations

    Sep 8, 2026Priyanka Mary Mammen, Emil Joswin, Srujananjali MedicherlaAI Agent MonitoringRepresentation Probing

    UMass Amherst · Independent Research

  19. When Decodability Is Not Enough: Logical Validity Representations, Behavioral Dissociation, and Causal Tests in Language Models

    Sep 2, 2026Smitha Muthya Sudheendra, Jaideep SrivastavaLogical ReasoningRepresentation Probing

    University of Minnesota, Twin Cities

  20. What, Where, and How: Probing Spatiotemporal Representations in Video Foundation Models

    Sep 1, 2026Sharon S. Musa, Fereshteh Forghani, Harrish Thasarathan +3Video Representation LearningVision Foundation Models

    Lassonde School of Engineering, York University, Toronto, Canada · Vector Institute, Toronto, Canada · FAIR, Meta Superintelligence Labs +2

  21. Excess Separability: Nuisance-Controlled Residual-Stream Probing for Benchmark Contamination Detection

    Aug 12, 2026Florian BraunBenchmark ContaminationData Contamination in Language Models

  22. P3CA: Encoder-Agnostic Interpretation of Vision Foundation Model Embeddings via Spatial Probing

    Aug 10, 2026Amoon Jamzad, Dilakshan Srikanthan, Faranak Akbarifar +2Vision Foundation ModelsRepresentation Probing

    School of Computing, Queen’s University, Kingston, Ontario, K7L 2N8, Canada

  23. Support Operation Factorization: Compositional Readout of Frozen Vision Encoders under Controlled Interventions

    Aug 6, 2026Zhongyao Wang, Wanli Ouyang, Taoyong Cui +1Visual Representation LearningRepresentation Disentanglement

    Fudan University · MMLab, The Chinese University of Hong Kong · Department of Computer Science and Engineering, The Chinese University of Hong Kong

  24. Probing Character-level Transformers for the Spanish L-shaped Morphome

    Aug 4, 2026Akhilesh Kakolu Ramarao, Kevin Tang, Wiebke Petersen +1Transformer InterpretabilityTransformer

    Department of English Language and Linguistics, Heinrich Heine University Düsseldorf · Department of Linguistics, College of Liberal Arts and Sciences, University of Florida · Institute of Linguistics and Information Science, Heinrich Heine University Düsseldorf +1

  25. ChaosProbe: A Neurochaotic Lens on Frozen Transformer Input-Embedding Spaces

    Aug 3, 2026Kunal Kumar Pant, Nithin NagarajLanguage Model FingerprintingRepresentation Probing

    Department of Electronics and Communication Engineering (ECE) Amrita Vishwa Vidyapeetham Kollam, Kerala, India · Complex Systems Programme National Institute of Advanced Studies (NIAS) IISc Campus, Bengaluru Karnataka, India

  26. Probing the 3D Object-Level Understanding of Pre-Trained Detection Transformers

    Aug 2, 2026Robin Kim, Colin Samplawski, Benjamin M. MarlinDetection TransformerRepresentation Probing

    UMass Amherst, Amherst, MA, USA · SRI International, Menlo Park, CA, USA

  27. Sparse Concept Channels in Frozen 3D CT Vision Encoders

    Jul 23, 2026Farhad Nooralahzadeh, Lea Bogensperger, Christian Bluethgen +1Radiology Report Generation3D Medical Imaging

    Department of Quantitative Biomedicine, University of Zurich, Switzerland · Center for Artificial Intelligence in Medicine and Imaging, Stanford University, USA

  28. Screening of Biosecurity Features in Metagenomic Data with Evo 2 Probes

    Jul 15, 2026Jeremy Guntoro, Alexander Dack, Dylan Danno +3Representation ProbingGenomics

    John Innes Centre · AIxBio Hackathon contestant · Department of Bioengineering, Imperial College London +1

  29. Analysis-by-Proxy: Localization Signals in VLMs Operating as Condition Encoders

    Jul 7, 2026Yoav Baron, Sara Dorfman, Roni Paiss +2Visual Place RecognitionVLM Interpretability

    Tel Aviv University, Tel Aviv, Israel · Google DeepMind.

  30. Probing Geospatial SSL Representations with Environmental Signals

    Jul 6, 2026Rohita Mocharla, Vishal M. PatelRepresentation GeometryGeospatial Representation Learning

    Johns Hopkins Applied Physics Laboratory · Johns Hopkins University · Laurel, MD +1