Mechanistic Interpretability

Ask about this topic

Overview

The field seeks to explain how a model’s internal computations produce particular outputs, rather than relying only on input-output behavior. Researchers use sparse autoencoders and transcoders to identify features, and probes, attribution graphs, activation patching, ablation, and steering to trace or test their roles. Studies examine language-model tasks such as factual recall, reasoning, and refusal, as well as vision, audio, and scientific models. A central challenge is that a signal can be easy to decode without being causally used, while circuit explanations and interventions can vary across prompts, models, and analysis choices.

Latest papers

CardsList

391 papers63 papers in the last four weeks, up 75% on the four weeks before. 0.4% of all new papers.

  1. LAWFUL: Law-Aligned Witness for Faithful Use of Latents

    Jul 26, 2026Kevin Chen, Kenneth W. Parker, Anish AroraTransformer InterpretabilityMechanistic Interpretability

    Computer Science & Engineering, The Ohio State University, Columbus, Ohio, USA · Independent Researcher, Waterford, Virginia, USA

  2. Toward Mechanistic Interpretability of an AI Foundation Model Fine-Tuned for Atmospheric Chemistry

    Jul 22, 2026Jason Y. Hu, Ivan Higuera-Mendieta, Patrick Obin Sturm +1Air Quality PredictionMechanistic Interpretability

    Stanford University, Department of Computer Science, Stanford, CA, USA. · Stanford University, Department of Earth System Science, Stanford, CA, USA · University of Southern California, Department of Earth Sciences, Los Angeles, CA, USA +1

  3. Gotta Catch them all: the modes of Sycophancy

    Jul 22, 2026Shreyans Jain, Alexandra Yost, Amirali AbdullahLLM SycophancyLLM Interpretability

    Thoughtworks · Southern Utah University (SUU)

  4. Reading and Steering Representations of Materials-Science Mechanisms in an Open-Weight Language Model

    Jul 22, 2026Markus J. BuehlerMechanistic InterpretabilityScientific Reasoning in Language Models

    Department of Civil and Environmental Engineering · Department of Mechanical Engineering · Schwarzman College of Computing +1

  5. A Multi-Dimensional Evaluation of Explainability in Media Bias Detection

    Jul 22, 2026Ting Chen, Raina Zhang, Benjamin M. Ampel +1Attention Head AnalysisMechanistic Interpretability

    Carnegie Mellon University · Indiana University · Georgia State University

  6. CircuitKIT : Circuit Discovery, Evaluation, and Application Toolkit for Mechanistic Interpretability

    Jul 21, 2026Pratinav Seth, Hem Gosalia, Aditya Kasliwal +1Circuit DiscoveryMechanistic Interpretability

    Lexsi Labs

  7. Semantic Primes as Explanans for Emotion in Large Language Models

    Jul 21, 2026Frank XingLLM InterpretabilityMechanistic Interpretability

    University of Reading

  8. Measuring Monosemanticity in Sparse Autoencoders via Latent Activation Coherence

    Jul 20, 2026Katarzyna Filus, Sebastian PokucińskiRepresentation DisentanglementSparse Autoencoders

    Institute of Theoretical and Applied Informatics, Polish Academy of Sciences, Bałtycka 5, 44-100 Gliwice, Poland · Department of Applied Informatics, Silesian University of Technology, Akademicka 16, 44-100 Gliwice, Poland

  9. Persistent Sparse Autoencoders: Learning Feature-Specific Timescales in Language Model Representations

    Jul 19, 2026Haoyan Luo, Mateo Espinosa Zarlenga, Mateja JamnikSparse AutoencodersMechanistic Interpretability

    University of Cambridge · University of Oxford

  10. Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent

    Jul 17, 2026Sriram Balasubramanian, Soheil FeiziMechanistic InterpretabilityNeural Network Interpretability

    Department of Computer Science University of Maryland

  11. Toward a mechanistic understanding of inference in visual cortex and diffusion models

    Jul 17, 2026Zeyu Yun, Alexander Belsten, Dasheng Bi +3Mechanistic InterpretabilityVisual Neuroscience

    Dept. of Electrical Engineering and Computer Science, UC Berkeley · Redwood Center for Theoretical Neuroscience, UC Berkeley · Herbert Wertheim School of Optometry and Vision Science, UC Berkeley +3

  12. Steering Robustness into World Action Models via Mechanistic Interpretability and Optimal Control

    Jul 16, 2026Jihoon Hong, Julian Skifstad, Qiyue Dai +2World Models for RoboticsAction-Conditioned World Models

    Georgia Institute of Technology

  13. Tracing LLM Behavior to the Training Data with Empirical Next-Token Distributions

    Jul 15, 2026Zachary IzzoLanguage Model PretrainingTransformer Interpretability

    NEC Labs America

  14. AIMO Interpretability Challenge

    Jul 15, 2026Michal Štefánik, Philipp Mondorf, Andreas Waldis +15Mathematical Reasoning BenchmarksLLM Interpretability

    National Institute of Informatics, Japan · Munich Center for Machine Learning / MaiNLP LMU · University of Tübingen +6

  15. Minionese: Comprehensive Benchmark and Mechanistic Study of Multilingual LLM Safety

    Jul 11, 2026Chigozirim Ifebi, Brent Kong, Ayushi MehrotraLLM Safety BenchmarksMultilingual Language Model Evaluation

    California Institute of Technology.

  16. Training, Reading, and Editing Legible Transformers

    Jul 9, 2026Mark OskinTransformer InterpretabilityTransformer FFNs

    School of Computer Science and Engineering University of Washington

  17. Cross-seed explainability using Procrustes-conditioned Joint End-to-end Top-K Sparse Autoencoders

    Jul 9, 2026Bendegúz Váradi, Zoltán KmettyOrthogonal Procrustes AlignmentTransformer Interpretability

    Centre for Social Sciences, CSS-RECENS Research Group, Budapest 1097, Hungary. · Department of Sociology, Faculty of Social Sciences, Eötvös Loránd University, Budapest, Hungary.

  18. Towards Mechanistically Understanding Why Memorized Knowledge Fails to Generalize in Large Language Model Finetuning

    Jul 9, 2026Lu Dai, Ziyang Rao, Yili Wang +3Knowledge DistillationLLM Fine-Tuning

    HKUST · HKUST(GZ)

  19. Certified Interventional Fidelity: Anytime-Valid, Adaptive Evaluation of Causal Claims in Mechanistic Interpretability

    Jul 9, 2026Amir AsiaeeMechanistic InterpretabilityCausal Interventions in Language Models

    Department of Biostatistics, Vanderbilt University Medical Center, Nashville, TN 37232, USA

  20. Mechanistic Interpretability of LLM Jailbreaks via Internal Attribution Graphs

    Jul 8, 2026Anupam Wagle, Ifrat Ikhtear Uddin, Chaowei Zhang +1LLM InterpretabilityAdversarial Attacks on LLMs

    Department of Computer Science, University of South Dakota, USA · School of Information and Artificial Intelligence, Yangzhou University, Yangzhou, China

  21. Reliable mechanistic operator recovery with biologically-informed neural networks: principles for architecture and optimisation design

    Jul 8, 2026Rebecca M. Crossley, Yuan Yin, Sarah L. Waters +1Neural Network OptimizationMechanistic Interpretability

    Mathematical Institute, University of Oxford, Oxford, United Kingdom, OX2 6GG

  22. Mechanistic Interpretability for Neural Networks: Circuits, Sparse Features and Symbolic Reasoning

    Jul 8, 2026Pranav Sawant, Jakub KrejčíTransformer InterpretabilityMechanistic Interpretability

    The University of Texas at Dallas 800 W Campbell Rd, Richardson, TX 75080, United States · VSB—Technical University of Ostrava 708 00, Ostrava, Czech Republic

  23. Covert Trait Propagation Is Representation Alignment: Mechanistic Evidence from Hidden-Channel Distillation

    Jul 5, 2026Kargi Chauhan, Aditya ShahMechanistic InterpretabilityKnowledge Distillation

    University of California, Santa Cruz · 2Google

  24. Pathways of Visual Information Flow in Vision-Language Models

    Jul 3, 2026Israfel Salazar, Stella Frank, Dan Oneata +2Vision-Language ModelsVLM Interpretability

    University of Copenhagen · Technical University of Denmark · POLITEHNICA Bucharest +1

  25. Individual Parameters in Weight-Sparse Transformers Appear Interpretable

    Jul 3, 2026Arnau Marin-Llobet, Stefan HeimersheimTransformer InterpretabilityTransformer

    School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138

  26. Towards Robustness against Typographic Attack with Training-free Concept Localization

    Jul 2, 2026Bohan Liu, Wenqian Ye, Guangzhi Xiong +3VLM RobustnessAdversarial Attacks on VLMs

    University of Virginia, Charlottesville VA 22903, USA

  27. Conditional Co-Ablation: Recovering Self-Repair Backups in Transformer Circuits

    Jul 2, 2026Zhiren Gong, He Lu, Tiantong Wang +6Circuit DiscoveryMechanistic Interpretability

    College of Computing and Data Science, Nanyang Technological University, Singapore · Interdisciplinary Graduate Programme, Nanyang Technological University, Singapore · School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore

  28. Expander Sparse Autoencoders: Parameter-Efficient Dictionaries for Mechanistic Interpretability

    Jul 2, 2026Rodrigo Mendoza-SmithSparse RecoveryActivation Sparsity

    Independent Researcher.

  29. Mechanistic Interpretability and Causal Feature Steering of Neural Quantum States via Sparse Autoencoders

    Jul 1, 2026Zihao Qi, Christopher EarlsNeural Quantum StatesMechanistic Interpretability

    Department of Physics, Cornell University, Ithaca, NY 14853, USA. · Center for Applied Mathematics, Cornell University, Ithaca, NY 14853, USA.

  30. The Model Organism Lottery: Model Organism Interpretability Strongly Depends on Training Methodology

    Jul 1, 2026Andrzej Szablewski, Gabriel Konar-Steenberg, Raffaello Fornasiere +2Mechanistic InterpretabilityLanguage Model Post-Training

    1LASR Labs