{"total":79,"rows":[{"id":"2404.14082","title":"Mechanistic Interpretability for AI Safety -- A Review","created":"2024-04-22","cited_by_count":422,"area":"mechinterp"},{"id":"2404.03592","title":"ReFT: Representation Finetuning for Language Models","created":"2024-04-04","cited_by_count":197,"area":null},{"id":"2410.20526","title":"Llama Scope: Extracting Millions of Features from Llama-3.1-8B with Sparse Autoencoders","created":"2024-10-27","cited_by_count":143,"area":"mechinterp"},{"id":"2503.09532","title":"SAEBench: A Comprehensive Benchmark for Sparse Autoencoders in Language Model Interpretability","created":"2025-03-12","cited_by_count":124,"area":"mechinterp"},{"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":"2502.16681","title":"Are Sparse Autoencoders Useful? A Case Study in Sparse Probing","created":"2025-02-23","cited_by_count":115,"area":"mechinterp"},{"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":"2504.13151","title":"MIB: A Mechanistic Interpretability Benchmark","created":"2025-04-17","cited_by_count":56,"area":"mechinterp"},{"id":"2407.10223","title":"On Large Language Model Continual Unlearning","created":"2024-07-14","cited_by_count":52,"area":"safety"},{"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":"2505.13898","title":"Do Language Models Use Their Depth Efficiently?","created":"2025-05-20","cited_by_count":50,"area":null},{"id":"2408.10920","title":"Recurrent Neural Networks Learn to Store and Generate Sequences using Non-Linear Representations","created":"2024-08-20","cited_by_count":47,"area":null},{"id":"2501.08319","title":"Enhancing Automated Interpretability with Output-Centric Feature Descriptions","created":"2025-01-14","cited_by_count":41,"area":null},{"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":"2502.17516","title":"A Survey on Mechanistic Interpretability for Multi-Modal Foundation Models","created":"2025-02-22","cited_by_count":34,"area":"mllm"},{"id":"2406.11614","title":"Intrinsic Test of Unlearning Using Parametric Knowledge Traces","created":"2024-06-17","cited_by_count":29,"area":"safety"},{"id":"2605.05115","title":"Manifold Steering Reveals the Shared Geometry of Neural Network Representation and Behavior","created":"2026-05-06","cited_by_count":26,"area":null},{"id":"2407.14494","title":"InterpBench: Semi-Synthetic Transformers for Evaluating Mechanistic Interpretability Techniques","created":"2024-07-19","cited_by_count":25,"area":"mechinterp"},{"id":"2408.01416","title":"The Quest for the Right Mediator: Surveying Mechanistic Interpretability Through the Lens of Causal Mediation Analysis","created":"2024-08-02","cited_by_count":24,"area":"mechinterp"},{"id":"2409.04185","title":"Residual Stream Analysis with Multi-Layer SAEs","created":"2024-09-06","cited_by_count":19,"area":"mechinterp"},{"id":"2602.14111","title":"Sanity Checks for Sparse Autoencoders: Do SAEs Beat Random Baselines?","created":"2026-02-15","cited_by_count":14,"area":"mechinterp"},{"id":"2505.16004","title":"Evaluating Adversarial Robustness of Concept Representations in Sparse Autoencoders","created":"2025-05-21","cited_by_count":12,"area":"mechinterp"},{"id":"2509.13281","title":"RepIt: Steering Language Models with Concept-Specific Refusal Vectors","created":"2025-09-16","cited_by_count":12,"area":null},{"id":"2602.02464","title":"From Directions to Regions: Decomposing Activations in Language Models via Local Geometry","created":"2026-02-02","cited_by_count":12,"area":null},{"id":"2408.15510","title":"How Reliable are Causal Probing Interventions?","created":"2024-08-28","cited_by_count":11,"area":null},{"id":"2410.22590","title":"Characterizing the Role of Similarity in the Property Inferences of Language Models","created":"2024-10-29","cited_by_count":11,"area":null},{"id":"2505.11770","title":"Internal Causal Mechanisms Robustly Predict Language Model Out-of-Distribution Behaviors","created":"2025-05-17","cited_by_count":11,"area":null},{"id":"2505.22586","title":"Precise In-Parameter Concept Erasure in Large Language Models","created":"2025-05-28","cited_by_count":11,"area":null},{"id":"2508.11214","title":"How Causal Abstraction Underpins Computational Explanation","created":"2025-08-15","cited_by_count":11,"area":null},{"id":"2511.07772","title":"SALT: Steering Activations towards Leakage-free Thinking in Chain of Thought","created":"2025-11-11","cited_by_count":11,"area":"lr"},{"id":"2411.02124","title":"Adaptive Sparse Allocation with Mutual Choice & Feature Choice Sparse Autoencoders","created":"2024-11-04","cited_by_count":9,"area":"mechinterp"},{"id":"2503.10894","title":"HyperDAS: Towards Automating Mechanistic Interpretability with Hypernetworks","created":"2025-03-13","cited_by_count":9,"area":"mechinterp"},{"id":"2503.03032","title":"SAFE: A Sparse Autoencoder-Based Framework for Robust Query Enrichment and Hallucination Mitigation in LLMs","created":"2025-03-04","cited_by_count":8,"area":"mechinterp"},{"id":"2508.01916","title":"Decomposing Representation Space into Interpretable Subspaces with Unsupervised Learning","created":"2025-08-03","cited_by_count":8,"area":"mechinterp"},{"id":"2509.22033","title":"OrtSAE: Orthogonal Sparse Autoencoders Uncover Atomic Features","created":"2025-09-26","cited_by_count":8,"area":"mechinterp"},{"id":"2602.05027","title":"AudioSAE: Towards Understanding of Audio-Processing Models with Sparse AutoEncoders","created":"2026-02-04","cited_by_count":8,"area":"mechinterp"},{"id":"2511.04638","title":"Addressing divergent representations from causal interventions on neural networks","created":"2025-11-06","cited_by_count":7,"area":"mechinterp"},{"id":"2512.15134","title":"From Isolation to Entanglement: When Do Interpretability Methods Identify and Disentangle Known Concepts?","created":"2025-12-17","cited_by_count":7,"area":null},{"id":"2602.11881","title":"From Atoms to Trees: Building a Structured Feature Forest with Hierarchical Sparse Autoencoders","created":"2026-02-12","cited_by_count":6,"area":"mechinterp"},{"id":"2409.13714","title":"TracrBench: Generating Interpretability Testbeds with Large Language Models","created":"2024-09-07","cited_by_count":5,"area":null},{"id":"2506.10920","title":"Constructing Interpretable Features from Compositional Neuron Groups","created":"2025-06-12","cited_by_count":5,"area":"mechinterp"},{"id":"2506.06686","title":"Learning Distribution-Wise Control in Representation Space for Language Models","created":"2025-06-07","cited_by_count":4,"area":null},{"id":"2509.23717","title":"Measuring Sparse Autoencoder Feature Sensitivity","created":"2025-09-28","cited_by_count":4,"area":"mechinterp"},{"id":"2604.11061","title":"Pando: Do Interpretability Methods Work When Models Won't Explain Themselves?","created":"2026-04-13","cited_by_count":4,"area":"mechinterp"},{"id":"2605.18229","title":"Are Sparse Autoencoder Benchmarks Reliable?","created":"2026-05-18","cited_by_count":4,"area":"mechinterp"},{"id":"2507.10085","title":"Enhancing Chain-of-Thought Reasoning with Critical Representation Fine-tuning","created":"2025-07-14","cited_by_count":3,"area":"lr"},{"id":"2602.02467","title":"Indications of Belief-Guided Agency and Meta-Cognitive Monitoring in Large Language Models","created":"2026-02-02","cited_by_count":3,"area":null},{"id":"2410.07456","title":"SAGE: Scalable Ground Truth Evaluations for Large Sparse Autoencoders","created":"2024-10-09","cited_by_count":2,"area":"mechinterp"},{"id":"2505.10375","title":"Are Sparse Autoencoders Useful for Java Function Bug Detection?","created":"2025-05-15","cited_by_count":2,"area":"mechinterp"},{"id":"2506.07406","title":"InverseScope: Scalable Activation Inversion for Interpreting Large Language Models","created":"2025-06-09","cited_by_count":2,"area":"mechinterp"},{"id":"2507.05158","title":"Steering Information Utility in Key-Value Memory for Language Model Post-Training","created":"2025-07-07","cited_by_count":2,"area":null},{"id":"2507.05387","title":"The Generalization Ridge: Information Flow in Natural Language Generation","created":"2025-07-07","cited_by_count":2,"area":null},{"id":"2605.01381","title":"A framework for analyzing concept representations in neural models","created":"2026-05-02","cited_by_count":2,"area":null},{"id":"2605.23035","title":"Sparse Autoencoders Map Brain-LLM Alignment onto Cortical Semantic Topography","created":"2026-05-21","cited_by_count":2,"area":"alignment"},{"id":"2410.09349","title":"Inference and Verbalization Functions During In-Context Learning","created":"2024-10-12","cited_by_count":1,"area":null},{"id":"2502.03618","title":"The Logical Implication Steering Method for Conditional Interventions on Transformer Generation","created":"2025-02-05","cited_by_count":1,"area":"mechinterp"},{"id":"2509.24502","title":"SUIT: Knowledge Editing with Subspace-Aware Key-Value Mappings","created":"2025-09-29","cited_by_count":1,"area":null},{"id":"2511.18409","title":"Findings of the BlackboxNLP 2025 Shared Task: Localizing Circuits and Causal Variables in Language Models","created":"2025-11-23","cited_by_count":1,"area":"mechinterp"},{"id":"2605.02234","title":"Bucketing the Good Apples: A Method for Diagnosing and Improving Causal Abstraction","created":"2026-05-04","cited_by_count":1,"area":null},{"id":"2605.24942","title":"Riemannian-Manifold Steering: Geometry-Aware Generative Autoencoders for Label-Free Steering","created":"2026-05-24","cited_by_count":1,"area":null},{"id":"2607.24645","title":"Sparse Autoencoders Encode Both Concepts and Functions: The Downstream Geometry of Feature Effects","created":"2026-07-27","cited_by_count":1,"area":"mechinterp"},{"id":"2608.15507","title":"Do Language Models Consistently Encode the Current Year?","created":"2026-08-16","cited_by_count":1,"area":"lr"},{"id":"2410.16484","title":"Identifying Sub-networks in Neural Networks via Functionally Similar Representations","created":"2024-10-21","cited_by_count":0,"area":null},{"id":"2508.06592","title":"Towards Integrated Alignment","created":"2025-08-08","cited_by_count":0,"area":"alignment"},{"id":"2512.18505","title":"Teaching and Critiquing Conceptualization and Operationalization in NLP","created":"2025-12-20","cited_by_count":0,"area":null},{"id":"2602.15730","title":"Causal Effect Estimation with Latent Textual Treatments","created":"2026-02-17","cited_by_count":0,"area":null},{"id":"2605.14749","title":"Non-linear Interventions on Large Language Models","created":"2026-05-14","cited_by_count":0,"area":"mechinterp"},{"id":"2607.00267","title":"Validating Causal Abstraction Metrics on Simulated Complex Systems","created":"2026-06-30","cited_by_count":0,"area":null},{"id":"2607.10248","title":"One mechanism for many mental spaces: a shared router over a value slot in language models","created":"2026-07-11","cited_by_count":0,"area":null},{"id":"2607.13899","title":"AIMO Interpretability Challenge","created":"2026-07-15","cited_by_count":0,"area":"lr"},{"id":"2608.13337","title":"Where You Measure Decides What You Measure: Position Selection in Ablation-Based SAE Evaluation","created":"2026-08-13","cited_by_count":0,"area":"mechinterp"},{"id":"2609.08322","title":"Tracing Stereotypes from Representation to Output in Multilingual LLMs","created":"2026-09-08","cited_by_count":0,"area":"mechinterp"},{"id":"2609.08585","title":"Limitations of Automated Simulatability: LLM Simulators Can Bypass Explanations","created":"2026-09-08","cited_by_count":0,"area":null},{"id":"2609.09113","title":"SAEScientist-Bench: Can AI Agents Conduct Autonomous SAE Interpretability Research?","created":"2026-09-08","cited_by_count":0,"area":"mechinterp"},{"id":"2609.29328","title":"Grammatical \"grandmother neurons\" are rare in LLMs","created":"2026-09-24","cited_by_count":0,"area":null},{"id":"2609.32907","title":"Logical subspace in LLMs","created":"2026-09-26","cited_by_count":0,"area":null},{"id":"2609.35272","title":"When Words Speak Louder than Images: Towards Understanding Language Bias in Vision-Language Models","created":"2026-09-28","cited_by_count":0,"area":"mllm"},{"id":"2610.07270","title":"Does the Model Use the Feature? Separating Steering from Mechanism in LLMs","created":"2026-10-05","cited_by_count":0,"area":null},{"id":"2610.19359","title":"Steering Vectors for Competing Ethical Frameworks Are Interchangeable: A Competing-Target Test for Activation Steering","created":"2026-10-14","cited_by_count":0,"area":"mechinterp"}],"next_cursor":null}