{"total":235,"rows":[{"id":"2404.14082","title":"Mechanistic Interpretability for AI Safety -- A Review","created":"2024-04-22","cited_by_count":422,"area":"mechinterp"},{"id":"2406.01506","title":"The Geometry of Categorical and Hierarchical Concepts in Large Language Models","created":"2024-06-03","cited_by_count":159,"area":null},{"id":"2411.16724","title":"Devils in Middle Layers of Large Vision-Language Models: Interpreting, Detecting and Mitigating Object Hallucinations via Attention Lens","created":"2024-11-23","cited_by_count":143,"area":"mllm"},{"id":"2407.02646","title":"A Practical Review of Mechanistic Interpretability for Transformer-Based Language Models","created":"2024-07-02","cited_by_count":118,"area":"mechinterp"},{"id":"2410.13928","title":"Automatically Interpreting Millions of Features in Large Language Models","created":"2024-10-17","cited_by_count":113,"area":"mechinterp"},{"id":"2410.02762","title":"Interpreting and Editing Vision-Language Representations to Mitigate Hallucinations","created":"2024-10-03","cited_by_count":104,"area":"mllm"},{"id":"2601.07372","title":"Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models","created":"2026-01-12","cited_by_count":86,"area":"moe"},{"id":"2405.00208","title":"A Primer on the Inner Workings of Transformer-based Language Models","created":"2024-04-30","cited_by_count":85,"area":null},{"id":"2402.17700","title":"RAVEL: Evaluating Interpretability Methods on Disentangling Language Model Representations","created":"2024-02-27","cited_by_count":79,"area":"mechinterp"},{"id":"2406.12775","title":"Hopping Too Late: Exploring the Limitations of Large Language Models on Multi-Hop Queries","created":"2024-06-18","cited_by_count":78,"area":null},{"id":"2502.15603","title":"Do Multilingual LLMs Think In English?","created":"2025-02-21","cited_by_count":78,"area":"lr"},{"id":"2607.15495","title":"Verbalizable Representations Form a Global Workspace in Language Models","created":"2026-07-16","cited_by_count":75,"area":null},{"id":"2411.04986","title":"The Semantic Hub Hypothesis: Language Models Share Semantic Representations Across Languages and Modalities","created":"2024-11-07","cited_by_count":69,"area":null},{"id":"2403.10949","title":"SelfIE: Self-Interpretation of Large Language Model Embeddings","created":"2024-03-16","cited_by_count":67,"area":null},{"id":"2405.17969","title":"Knowledge Circuits in Pretrained Transformers","created":"2024-05-28","cited_by_count":60,"area":"mechinterp"},{"id":"2406.01563","title":"LoFiT: Localized Fine-tuning on LLM Representations","created":"2024-06-03","cited_by_count":57,"area":null},{"id":"2407.17817","title":"Demystifying Verbatim Memorization in Large Language Models","created":"2024-07-25","cited_by_count":56,"area":null},{"id":"2512.15674","title":"Activation Oracles: Training and Evaluating LLMs as General-Purpose Activation Explainers","created":"2025-12-17","cited_by_count":52,"area":null},{"id":"2407.15017","title":"Knowledge Mechanisms in Large Language Models: A Survey and Perspective","created":"2024-07-22","cited_by_count":51,"area":null},{"id":"2412.08686","title":"LatentQA: Teaching LLMs to Decode Activations Into Natural Language","created":"2024-12-11","cited_by_count":51,"area":null},{"id":"2402.09631","title":"Representation Surgery: Theory and Practice of Affine Steering","created":"2024-02-15","cited_by_count":50,"area":null},{"id":"2403.07175","title":"Rebuilding ROME : Resolving Model Collapse during Sequential Model Editing","created":"2024-03-11","cited_by_count":50,"area":"synth"},{"id":"2410.05864","title":"From Tokens to Words: On the Inner Lexicon of LLMs","created":"2024-10-08","cited_by_count":48,"area":null},{"id":"2402.14328","title":"Understanding and Patching Compositional Reasoning in LLMs","created":"2024-02-22","cited_by_count":46,"area":"lr"},{"id":"2411.08745","title":"Separating Tongue from Thought: Activation Patching Reveals Language-Agnostic Concept Representations in Transformers","created":"2024-11-13","cited_by_count":44,"area":"lr"},{"id":"2406.12094","title":"Who's asking? User personas and the mechanics of latent misalignment","created":"2024-06-17","cited_by_count":40,"area":"safety"},{"id":"2406.12673","title":"Estimating Knowledge in Large Language Models Without Generating a Single Token","created":"2024-06-18","cited_by_count":40,"area":null},{"id":"2403.18167","title":"Mechanistic Understanding and Mitigation of Language Model Non-Factual Hallucinations","created":"2024-03-27","cited_by_count":38,"area":null},{"id":"2410.06331","title":"Locate-then-edit for Multi-hop Factual Recall under Knowledge Editing","created":"2024-10-08","cited_by_count":37,"area":null},{"id":"2410.22366","title":"One-Step is Enough: Sparse Autoencoders for Text-to-Image Diffusion Models","created":"2024-10-28","cited_by_count":33,"area":"diffusion"},{"id":"2402.12865","title":"Backward Lens: Projecting Language Model Gradients into the Vocabulary Space","created":"2024-02-20","cited_by_count":32,"area":null},{"id":"2411.04430","title":"Towards Unifying Interpretability and Control: Evaluation via Intervention","created":"2024-11-07","cited_by_count":32,"area":"mechinterp"},{"id":"2505.17127","title":"Pixels Versus Priors: Controlling Knowledge Priors in Vision-Language Models through Visual Counterfacts","created":"2025-05-21","cited_by_count":32,"area":"mllm"},{"id":"2511.08579","title":"Training Language Models to Explain Their Own Computations","created":"2025-11-11","cited_by_count":30,"area":null},{"id":"2509.12886","title":"The LLM Already Knows: Estimating LLM-Perceived Question Difficulty via Hidden Representations","created":"2025-09-16","cited_by_count":29,"area":null},{"id":"2411.16679","title":"Do Large Language Models Perform Latent Multi-Hop Reasoning without Exploiting Shortcuts?","created":"2024-11-25","cited_by_count":28,"area":"lr"},{"id":"2510.13900","title":"Narrow Finetuning Leaves Clearly Readable Traces in Activation Differences","created":"2025-10-14","cited_by_count":28,"area":null},{"id":"2502.05489","title":"Mechanistic Interpretability of Emotion Inference in Large Language Models","created":"2025-02-08","cited_by_count":27,"area":"mechinterp"},{"id":"2407.19200","title":"On Behalf of the Stakeholders: Trends in NLP Model Interpretability in the Era of LLMs","created":"2024-07-27","cited_by_count":25,"area":"mechinterp"},{"id":"2409.03621","title":"Attend First, Consolidate Later: On the Importance of Attention in Different LLM Layers","created":"2024-09-05","cited_by_count":25,"area":null},{"id":"2504.03022","title":"The Dual-Route Model of Induction","created":"2025-04-03","cited_by_count":24,"area":null},{"id":"2411.07404","title":"Controllable Context Sensitivity and the Knob Behind It","created":"2024-11-11","cited_by_count":23,"area":null},{"id":"2503.11667","title":"LogitLens4LLMs: Extending Logit Lens Analysis to Modern Large Language Models","created":"2025-02-24","cited_by_count":23,"area":"mechinterp"},{"id":"2407.07342","title":"Multilingual Blending: LLM Safety Alignment Evaluation with Language Mixture","created":"2024-07-10","cited_by_count":22,"area":"alignment"},{"id":"2502.17601","title":"Representation Engineering for Large-Language Models: Survey and Research Challenges","created":"2025-02-24","cited_by_count":21,"area":"mechinterp"},{"id":"2505.13141","title":"Language-Specific Latent Process Hinders Cross-Lingual Performance","created":"2025-05-19","cited_by_count":21,"area":null},{"id":"2505.20546","title":"Paths Not Taken: Understanding and Mending the Multilingual Factual Recall Pipeline","created":"2025-05-26","cited_by_count":21,"area":null},{"id":"2602.08169","title":"Spherical Steering: Geometry-Aware Activation Rotation for Language Models","created":"2026-02-09","cited_by_count":21,"area":null},{"id":"2406.12288","title":"An Investigation of Neuron Activation as a Unified Lens to Explain Chain-of-Thought Eliciting Arithmetic Reasoning of LLMs","created":"2024-06-18","cited_by_count":20,"area":"lr"},{"id":"2504.10191","title":"Localized Cultural Knowledge is Conserved and Controllable in Large Language Models","created":"2025-04-14","cited_by_count":20,"area":"wm"},{"id":"2405.20778","title":"Improved Generation of Adversarial Examples Against Safety-aligned LLMs","created":"2024-05-28","cited_by_count":18,"area":"safety"},{"id":"2506.05140","title":"AudioLens: A Closer Look at Auditory Attribute Perception of Large Audio-Language Models","created":"2025-06-05","cited_by_count":18,"area":null},{"id":"2408.03573","title":"AcTracer: Active Testing of Large Language Model via Multi-Stage Sampling","created":"2024-08-07","cited_by_count":17,"area":null},{"id":"2412.11965","title":"Inferring Functionality of Attention Heads from their Parameters","created":"2024-12-16","cited_by_count":17,"area":null},{"id":"2505.24858","title":"MetaFaith: Faithful Natural Language Uncertainty Expression in LLMs","created":"2025-05-30","cited_by_count":16,"area":null},{"id":"2506.00253","title":"Aligned but Blind: Alignment Increases Implicit Bias by Reducing Awareness of Race","created":"2025-05-30","cited_by_count":15,"area":"alignment"},{"id":"2602.00462","title":"LatentLens: Revealing Highly Interpretable Visual Tokens in LLMs","created":"2026-01-31","cited_by_count":15,"area":"mllm"},{"id":"2604.25783","title":"Subliminal Steering: Stronger Encoding of Hidden Signals","created":"2026-04-28","cited_by_count":14,"area":"mechinterp"},{"id":"2409.05771","title":"Evidence from fMRI Supports a Two-Phase Abstraction Process in Language Models","created":"2024-09-09","cited_by_count":13,"area":null},{"id":"2412.04139","title":"Monet: Mixture of Monosemantic Experts for Transformers","created":"2024-12-05","cited_by_count":13,"area":"mechinterp"},{"id":"2502.07424","title":"RomanLens: The Role Of Latent Romanization In Multilinguality In LLMs","created":"2025-02-11","cited_by_count":12,"area":"mechinterp"},{"id":"2605.26242","title":"Can LLMs Introspect? A Reality Check","created":"2026-05-25","cited_by_count":12,"area":null},{"id":"2412.01113","title":"LLMs Faithfully and Iteratively Compute Answers During CoT: A Systematic Analysis With Multi-step Arithmetics","created":"2024-12-02","cited_by_count":11,"area":"lr"},{"id":"2503.11302","title":"Are formal and functional linguistic mechanisms dissociated in language models?","created":"2025-03-14","cited_by_count":11,"area":null},{"id":"2510.05092","title":"Learning to Interpret Weight Differences in Language Models","created":"2025-10-06","cited_by_count":11,"area":null},{"id":"2502.04667","title":"Compositional Generalization from Learned Skills via CoT Training: A Theoretical and Structural Analysis for Reasoning","created":"2025-02-07","cited_by_count":10,"area":"lr"},{"id":"2506.05451","title":"Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety","created":"2025-06-05","cited_by_count":10,"area":"safety"},{"id":"2509.13316","title":"Do Activation Verbalization Methods Convey Privileged Information?","created":"2025-09-16","cited_by_count":10,"area":"mechinterp"},{"id":"2605.06510","title":"Is One Layer Enough? Understanding Inference Dynamics in Tabular Foundation Models","created":"2026-05-07","cited_by_count":10,"area":null},{"id":"2404.05971","title":"Does Transformer Interpretability Transfer to RNNs?","created":"2024-04-09","cited_by_count":9,"area":"ssm"},{"id":"2409.06328","title":"Extracting Paragraphs from LLM Token Activations","created":"2024-09-10","cited_by_count":9,"area":null},{"id":"2503.10894","title":"HyperDAS: Towards Automating Mechanistic Interpretability with Hypernetworks","created":"2025-03-13","cited_by_count":9,"area":"mechinterp"},{"id":"2503.16356","title":"CaKE: Circuit-aware Editing Enables Generalizable Knowledge Learners","created":"2025-03-20","cited_by_count":9,"area":null},{"id":"2410.13274","title":"Breaking Chains: Unraveling the Links in Multi-Hop Knowledge Unlearning","created":"2024-10-17","cited_by_count":8,"area":"safety"},{"id":"2504.09522","title":"How new data permeates LLM knowledge and how to dilute it","created":"2025-04-13","cited_by_count":8,"area":null},{"id":"2505.23701","title":"Can LLMs Reason Abstractly Over Math Word Problems Without CoT? Disentangling Abstract Formulation From Arithmetic Computation","created":"2025-05-29","cited_by_count":8,"area":"lr"},{"id":"2510.01685","title":"How Do Language Models Compose Functions?","created":"2025-10-02","cited_by_count":8,"area":null},{"id":"2602.04735","title":"From Data to Behavior: Predicting Unintended Model Behaviors Before Training","created":"2026-02-04","cited_by_count":8,"area":null},{"id":"2603.06495","title":"COLD-Steer: Steering Large Language Models via In-Context One-step Learning Dynamics","created":"2026-03-06","cited_by_count":8,"area":"mechinterp"},{"id":"2405.17653","title":"InversionView: A General-Purpose Method for Reading Information from Neural Activations","created":"2024-05-27","cited_by_count":7,"area":null},{"id":"2505.16477","title":"Advancing the Scientific Method with Large Language Models: From Hypothesis to Discovery","created":"2025-05-22","cited_by_count":7,"area":null},{"id":"2511.06571","title":"Rep2Text: Decoding Full Text from a Single LLM Token Representation","created":"2025-11-09","cited_by_count":7,"area":null},{"id":"2608.02657","title":"Your Agentic LLMs Secretly Encode Indirect Prompt-Injection Exposure in Hidden States","created":"2026-08-01","cited_by_count":7,"area":"agents"},{"id":"2502.12055","title":"Designing Role Vectors to Improve LLM Inference Behaviour","created":"2025-02-17","cited_by_count":6,"area":null},{"id":"2505.15784","title":"Large Language Models as Computable Approximations to Solomonoff Induction","created":"2025-05-21","cited_by_count":6,"area":null},{"id":"2505.20133","title":"Token Distillation: Attention-aware Input Embeddings For New Tokens","created":"2025-05-26","cited_by_count":6,"area":null},{"id":"2505.22630","title":"Stochastic Chameleons: Irrelevant Context Hallucinations Reveal Class-Based (Mis)Generalization in LLMs","created":"2025-05-28","cited_by_count":6,"area":null},{"id":"2510.04128","title":"Internal states before wait modulate reasoning patterns","created":"2025-10-05","cited_by_count":6,"area":"lr"},{"id":"2510.19555","title":"[De|Re]constructing VLMs' Reasoning in Counting","created":"2025-10-22","cited_by_count":6,"area":"lr"},{"id":"2601.03425","title":"The Illusion of Specialization: Unveiling the Domain-Invariant \"Standing Committee\" in Mixture-of-Experts Models","created":"2026-01-06","cited_by_count":6,"area":"moe"},{"id":"2604.12493","title":"Latent Planning Emerges with Scale","created":"2026-04-14","cited_by_count":6,"area":null},{"id":"2410.02472","title":"Meta-Models: An Architecture for Decoding LLM Behaviors Through Interpreted Embeddings and Natural Language","created":"2024-10-03","cited_by_count":5,"area":null},{"id":"2501.15754","title":"Weight-based Analysis of Detokenization in Language Models: Understanding the First Stage of Inference Without Inference","created":"2025-01-27","cited_by_count":5,"area":null},{"id":"2502.07813","title":"CryptoX : Compositional Reasoning Evaluation of Large Language Models","created":"2025-02-08","cited_by_count":5,"area":"lr"},{"id":"2508.05525","title":"The World According to LLMs: How Geographic Origin Influences LLMs' Entity Deduction Capabilities","created":"2025-08-07","cited_by_count":5,"area":null},{"id":"2512.03771","title":"In-Context Representation Hijacking","created":"2025-12-03","cited_by_count":5,"area":"safety"},{"id":"2512.19980","title":"Neuron-Guided Interpretation of Code LLMs: Where, Why, and How?","created":"2025-12-23","cited_by_count":5,"area":null},{"id":"2402.11355","title":"A Practical Method for Generating String Counterfactuals","created":"2024-02-17","cited_by_count":4,"area":null},{"id":"2404.03623","title":"Unveiling LLMs: The Evolution of Latent Representations in a Dynamic Knowledge Graph","created":"2024-04-04","cited_by_count":4,"area":null},{"id":"2410.11660","title":"Eliciting Textual Descriptions from Representations of Continuous Prompts","created":"2024-10-15","cited_by_count":4,"area":null}],"next_cursor":"100"}