LLM Interpretability

LLM: Large Language Model

Overview

Large language model interpretability asks what information models represent, how it moves through layers and components, and which computations cause particular answers or behaviors. Researchers use probes and representation geometry, sparse autoencoders, attribution and circuit tracing, activation patching or ablation, and steering to study factual recall, mathematical reasoning, hallucination, refusal, and in-context learning. These methods can reveal distributed circuits and layer-wise signals, and sometimes support targeted interventions without retraining. A central challenge is that information being decodable—or a direction being steerable—does not prove the model uses it; causal faithfulness, robustness across prompts and model families, and reliable evaluation remain open tensions.

Latest papers

CardsList

841 papers55 papers in the last four weeks, down 36% on the four weeks before. 0.4% of all new papers.

  1. Reasoning-Trace Analysis Using Large Language Models for Interpretable Q-Matrix Construction

    Nov 19, 2026Jihong Zhang, Zijing Jin, Mingfeng XueCognitive Diagnostic ModelsLLM Interpretability

  2. Understanding MoE Decisions Beyond Gradients

    Nov 19, 2026Zhen Xu, Zitao Zhang, Kevin Zhao +1LLM InterpretabilityMixture-of-Experts Interpretability

  3. The Depth of a Prediction: Where Language Models Resolve a Word Predicts Human Reading Time

    Nov 19, 2026Donato Riccio, Fabrizio MaturoLanguage ModelingLLM Interpretability

  4. GateWatch: Low-Overhead LLM Monitoring with Boolean Circuits

    Nov 18, 2026Firas Spanioli, Yaniv DavidLLM Safety MonitoringLanguage Model Probing

  5. Chain-of-Thought Hidden States Reveal Task-Solving Directions for Direct Answers

    Nov 18, 2026Houman Mehrafarin, Ioannis KonstasCoT ReasoningEfficient Language Model Reasoning

  6. A Mechanistic Analysis of In-Context Learning and Out-of-Context Hallucination

    Nov 18, 2026Tianxiang Chen, Jingyuan Zhou, Guoyao Yang +1Hallucination in Language ModelsIn-Context Learning

  7. Tracking Representations in LLM Reasoning Traces for Iterated Games

    Nov 17, 2026Chaipat Tirapongprasert, Rajarshi Das, Parijat DubeStrategic Reasoning Risks in Language ModelsCoT Faithfulness

  8. What MoE Routing Controls: Causal Interventions on Expert Trajectories

    Nov 17, 2026Mukund Athmaram, Manoj Rao, Liang MaMixture-of-Experts Language ModelsLanguage Model Steering

  9. Disambiguating Moral Circuits and Moral Features in Language Models: Moral content, normative judgment or cue words? Evidence from 14 models

    Nov 16, 2026Davide Liga, Liuwen Yu, Matteo MagniniMoral Reasoning in Language ModelsLLM Interpretability

  10. A Bag of Concepts? Testing Role-Filler Binding in the Verbalizable Workspace

    Nov 14, 2026Reya Dawlah, Elizabeth McMillin, Michael Shihong Zhang +4LLM InterpretabilityRepresentation Geometry in Language Models

  11. Geometrically Disentangling Concept Learning from the Language Modeling Loss

    Nov 13, 2026Yupei Wang, Neil Mallinar, Mikhail Belkin +1LLM InterpretabilityConcept Learning

  12. Dancing Tokens on the Vocabulary Network from LLM Generation

    Nov 13, 2026Yuan-Chao HuLLM InterpretabilityRepresentation Geometry in Language Models

  13. DECODE: Diagnosing Interpretability Failures in Automated Interpretation Process

    Nov 13, 2026Dohyun Kim, Jaeung Lee, Donghee Hong +1Automated InterpretabilityLLM Interpretability

  14. An Operationalization of Opaque Serial Depth

    Nov 12, 2026Nathan Sheffield, Alek Westover, Lukas Finnveden +3Implicit Reasoning in Language ModelsLLM Interpretability

  15. Tracing Knowledge and Reasoning in LLMs

    Nov 12, 2026Yang Wang, Xiao Wang, Yiqi Liu +4LLM EvaluationLanguage Model Reasoning Evaluation

  16. What Roles Do System Prompts Really Assign? Validating and Steering Occupational Role Vectors in LLMs

    Nov 11, 2026Jose L. Sampedro Mazon, Ilmi Yoon, Nadim SaadActivation SteeringLLM Interpretability

  17. Differential Backpack Language Models: Sense Inhibition via Signed Contextualization Weights

    Nov 11, 2026Konstantinos Skianis, Giannis Nikolentzos, Polykarpos Meladianos +1Language ModelingLLM Interpretability

  18. MicroLens: Read, Trace, and Rewire the Residual Stream with Two Lenses

    Nov 9, 2026Nika Zahedi, Fangyi Zhu, Anna WuLLM InterpretabilityLanguage Model Introspection

  19. Linear Probes to Bayesian Manifolds: Geometry of Arithmetic in Large Language Models

    Nov 8, 2026Anshul Kumar, Neha Prasad, Manoj Kumar +2LLM InterpretabilityRepresentation Geometry in Language Models

  20. A Layer-wise Study of How Style Rewriting Breaks the Calibration of Zero-Shot LLM Fake-News Detectors

    Nov 8, 2026Prabhjot Singh, Praney Goyal, Raksha SharmaLanguage Model RobustnessLanguage Model Calibration

  21. SWEET: Mitigating Spurious Feature Reliance in LLM Reasoning via Weight Editing

    Nov 7, 2026Zhizhang Fu, Lucas Zhao, Michalis Korakakis +3Faithfulness of Language Model ReasoningLLM Interpretability

  22. Tracing Psychometric Inference in Large Language Models

    Nov 6, 2026Chih-Hao Hsu, Feng-Chun B. Chou, Pin-Hao A. ChenPersonality Modeling in Language ModelsLLM Interpretability