Mechanistic Interpretability

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

427 papers58 papers in the last four weeks, up 23% on the four weeks before. 0.4% of all new papers.

  1. Behavioral Compositionality of Sparse Autoencoder Representations

    Oct 15, 2026Haotian Ma, Ziye Mei, Frederic SalaSparse AutoencodersCompositional Representation Learning

  2. Do Emergent Misalignment Models Share a Circuit?

    Oct 14, 2026Muhammad Haseeb, Muhammad TahirEmergent Misalignment in Language ModelsCircuits in Language Models

  3. Interpreting Language Model Parameters

    Oct 14, 2026Lucius Bushnaq, Dan Braun, Oliver Clive-Griffin +5Mechanistic InterpretabilityTransformer Interpretability

  4. Sparse Autoencoders for Mechanistic Interpretability in NLP: A Critical Survey of Recoverability, Interpretability, and Utility

    Oct 14, 2026Salima Lamsiyah, Adelaide Danilov, Oleksandr Marchenko Breneur +1Sparse AutoencodersLLM Interpretability

  5. Multi-Stream Residual Computation in Diffusion Language Models

    Oct 13, 2026Ekaterina Alimaskina, Suren Kazarian, Gleb Molodtsov +1Diffusion Language ModelsMasked Diffusion Models

  6. What Do Sparse Autoencoder Feature Discovery Scores Identify? From Activations to Interventions

    Oct 11, 2026Yifeng Guo, Yanran Li, Xiang ChenSparse AutoencodersMechanistic Interpretability

  7. FlyHack: Stress-Testing Mechanistic Claims in a Connectome-Constrained Visual Model

    Oct 9, 2026Aashish DhawanConnectome-Informed Neural NetworksMechanistic Interpretability

  8. How to train your model organism

    Oct 7, 2026Xilin Wang, David Bau, Byron C. WallaceMechanistic InterpretabilityExplainability Evaluation

  9. Fully Interpretable Minimal Transformers: From Geometry to Algorithm

    Oct 7, 2026Raneem Mahajne, Toviah MoldwinTransformer InterpretabilityMechanistic Interpretability

  10. How Do Agentic LLMs Decide to Call Tools? A Tool-Call Vector Shaped by Suppression

    Oct 7, 2026Xijie Gong, Tingxu Han, Jiahao Zhang +5LLM Tool UseLLM Decision-Making

  11. A Shortcut to Structure in AlphaFold 3

    Oct 6, 2026Jonathan Feldman, Jeffrey SkolnickProtein Structure PredictionMechanistic Interpretability

  12. Mechanistic Interpretability of Atmospheric Rivers in GraphCast

    Oct 6, 2026Madelyn Mathai, Timothy B. Higgins, Kevin M. Grise +2Mechanistic InterpretabilityWeather Forecasting

  13. Inference and learning in sparse autoencoders as natural gradient flow

    Oct 5, 2026Hadi Vafaii, Tejas Rao, David Chanin +6Gradient DescentSparse Autoencoders

  14. Does the Model Use the Feature? Separating Steering from Mechanism in LLMs

    Oct 5, 2026Tong Che, Yilong LiLLM InterpretabilityMechanistic Interpretability

  15. Tracing a Sparse Emotion-Control Circuit in LLM-Based Text-to-Speech

    Oct 4, 2026Hongfei Du, Jiacheng Shi, Yanfu Zhang +1Mechanistic InterpretabilityControllable Speech Generation

  16. A Graph-Based Inspection and Intervention Tool for Assessing Mechanistic Learning in PINNs

    Oct 4, 2026Adwait Patkhedkar, Alifaraz Lakhani, Prathmesh Mohite +1Physics-Informed MLMechanistic Interpretability

  17. ScopeSAE: Model-Scope Feature Discovery with Interpretable Layer Selection

    Oct 4, 2026Qingwen Zeng, Zehao Fu, Shuyu Meng +5Transformer InterpretabilitySparse Autoencoders

  18. Slaying the Hydra: Interaction-Aware Circuit Discovery in Language Models

    Oct 2, 2026Sankaran Vaidyanathan, Rafal Urbaniak, Emily Bunnapradist +2Circuits in Language ModelsMechanistic Interpretability

  19. Clinical Concept Centers in LLMs

    Oct 2, 2026Aishik Nagar, Abhishek Vaidyanathan, Arun-Kumar Kaliya-Perumal +2Mechanistic InterpretabilityCausal Interventions in Language Models

  20. From Isolated Feature to Orbits: Discovering Music Concepts via Multi-SAE Alignment

    Oct 1, 2026Liwei Lin, Gus XiaRepresentation LearningSparse Autoencoders

  21. Compact Language, Complex Model Shifts: How and Where Ambiguity and Underspecification Affect LLMs

    Sep 30, 2026Michaela Regneri, Nina Scheller, Sören LaueLLM TrainingMechanistic Interpretability

  22. Weights Read and Write Features: Scalable Parameter Decomposition Grounded in Activation Space

    Sep 29, 2026Tue M. Cao, Lisiane Pruinelli, My T. ThaiMechanistic InterpretabilityCircuits in Language Models

  23. When Models Don't Manipulate Manifolds: The Geometry of a Comparison Task

    Sep 29, 2026Sai Sumedh R. Hindupur, Hadas Orgad, Thomas Fel +1Numerical Reasoning in Language ModelsMechanistic Interpretability