{"total":455,"rows":[{"id":"2406.11717","title":"Refusal in Language Models Is Mediated by a Single Direction","created":"2024-06-17","cited_by_count":1032,"area":"safety"},{"id":"2309.00770","title":"Bias and Fairness in Large Language Models: A Survey","created":"2023-09-02","cited_by_count":639,"area":null},{"id":"2403.19647","title":"Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models","created":"2024-03-28","cited_by_count":435,"area":null},{"id":"2404.14082","title":"Mechanistic Interpretability for AI Safety -- A Review","created":"2024-04-22","cited_by_count":422,"area":"mechinterp"},{"id":"2103.00453","title":"Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP","created":"2021-02-28","cited_by_count":337,"area":null},{"id":"2106.13219","title":"Towards Understanding and Mitigating Social Biases in Language Models","created":"2021-06-24","cited_by_count":296,"area":null},{"id":"2306.03819","title":"LEACE: Perfect linear concept erasure in closed form","created":"2023-06-06","cited_by_count":267,"area":null},{"id":"2303.02536","title":"Finding Alignments Between Interpretable Causal Variables and Distributed Neural Representations","created":"2023-03-05","cited_by_count":209,"area":null},{"id":"2301.04709","title":"Causal Abstraction: A Theoretical Foundation for Mechanistic Interpretability","created":"2023-01-11","cited_by_count":205,"area":"mechinterp"},{"id":"2109.00725","title":"Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond","created":"2021-09-02","cited_by_count":178,"area":null},{"id":"2410.12877","title":"Improving Instruction-Following in Language Models through Activation Steering","created":"2024-10-15","cited_by_count":160,"area":"mechinterp"},{"id":"2309.17410","title":"Can Sensitive Information Be Deleted From LLMs? Objectives for Defending Against Extraction Attacks","created":"2023-09-29","cited_by_count":158,"area":null},{"id":"2304.00740","title":"Inspecting and Editing Knowledge Representations in Language Models","created":"2023-04-03","cited_by_count":141,"area":null},{"id":"2209.11326","title":"Towards Faithful Model Explanation in NLP: A Survey","created":"2022-09-22","cited_by_count":130,"area":null},{"id":"2108.12084","title":"Harms of Gender Exclusivity and Challenges in Non-Binary Representation in Language Technologies","created":"2021-08-27","cited_by_count":128,"area":null},{"id":"2112.00826","title":"Inducing Causal Structure for Interpretable Neural Networks","created":"2021-12-01","cited_by_count":110,"area":null},{"id":"2101.09523","title":"Debiasing Pre-trained Contextualised Embeddings","created":"2021-01-23","cited_by_count":107,"area":null},{"id":"2305.08809","title":"Interpretability at Scale: Identifying Causal Mechanisms in Alpaca","created":"2023-05-15","cited_by_count":103,"area":null},{"id":"2201.12091","title":"Linear Adversarial Concept Erasure","created":"2022-01-28","cited_by_count":101,"area":null},{"id":"2208.07998","title":"What Artificial Neural Networks Can Tell Us About Human Language Acquisition","created":"2022-08-17","cited_by_count":99,"area":null},{"id":"2110.08527","title":"An Empirical Survey of the Effectiveness of Debiasing Techniques for Pre-trained Language Models","created":"2021-10-16","cited_by_count":97,"area":null},{"id":"2308.10149","title":"A Survey on Fairness in Large Language Models","created":"2023-08-20","cited_by_count":95,"area":null},{"id":"2307.08678","title":"Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanations","created":"2023-07-17","cited_by_count":85,"area":null},{"id":"2404.00929","title":"A Survey on Multilingual Large Language Models: Corpora, Alignment, and Bias","created":"2024-04-01","cited_by_count":84,"area":null},{"id":"2108.03362","title":"On Measures of Biases and Harms in NLP","created":"2021-08-07","cited_by_count":81,"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":"2411.10915","title":"Bias in Large Language Models: Origin, Evaluation, and Mitigation","created":"2024-11-16","cited_by_count":77,"area":null},{"id":"2103.01378","title":"Contrastive Explanations for Model Interpretability","created":"2021-03-02","cited_by_count":75,"area":null},{"id":"2310.12560","title":"Fast Model Debias with Machine Unlearning","created":"2023-10-19","cited_by_count":72,"area":"safety"},{"id":"2506.11618","title":"Convergent Linear Representations of Emergent Misalignment","created":"2025-06-13","cited_by_count":71,"area":null},{"id":"2305.12798","title":"Word Embeddings Are Steers for Language Models","created":"2023-05-22","cited_by_count":70,"area":null},{"id":"2305.11840","title":"SeeGULL: A Stereotype Benchmark with Broad Geo-Cultural Coverage Leveraging Generative Models","created":"2023-05-19","cited_by_count":66,"area":null},{"id":"2409.20089","title":"Robust LLM safeguarding via refusal feature adversarial training","created":"2024-09-30","cited_by_count":64,"area":null},{"id":"2410.15999","title":"Steering Knowledge Selection Behaviours in LLMs via SAE-Based Representation Engineering","created":"2024-10-21","cited_by_count":63,"area":"mechinterp"},{"id":"2205.02526","title":"Theories of \"Gender\" in NLP Bias Research","created":"2022-05-05","cited_by_count":62,"area":null},{"id":"2105.06965","title":"Counterfactual Interventions Reveal the Causal Effect of Relative Clause Representations on Agreement Prediction","created":"2021-05-14","cited_by_count":59,"area":null},{"id":"2204.08831","title":"Probing for the Usage of Grammatical Number","created":"2022-04-19","cited_by_count":59,"area":null},{"id":"2404.01349","title":"Fairness in Large Language Models: A Taxonomic Survey","created":"2024-03-31","cited_by_count":58,"area":null},{"id":"2405.09719","title":"Spectral Editing of Activations for Large Language Model Alignment","created":"2024-05-15","cited_by_count":56,"area":"alignment"},{"id":"2203.09192","title":"Entropy-based Attention Regularization Frees Unintended Bias Mitigation from Lists","created":"2022-03-17","cited_by_count":55,"area":null},{"id":"2205.14140","title":"CEBaB: Estimating the Causal Effects of Real-World Concepts on NLP Model Behavior","created":"2022-05-27","cited_by_count":55,"area":null},{"id":"2310.11877","title":"The Curious Case of Hallucinatory (Un)answerability: Finding Truths in the Hidden States of Over-Confident Large Language Models","created":"2023-10-18","cited_by_count":55,"area":null},{"id":"2211.01528","title":"Fair and Optimal Classification via Post-Processing","created":"2022-11-03","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":"2505.13898","title":"Do Language Models Use Their Depth Efficiently?","created":"2025-05-20","cited_by_count":50,"area":null},{"id":"2101.10001","title":"Diverse Adversaries for Mitigating Bias in Training","created":"2021-01-25","cited_by_count":48,"area":null},{"id":"2201.12191","title":"Kernelized Concept Erasure","created":"2022-01-28","cited_by_count":48,"area":null},{"id":"2210.02938","title":"Debiasing isn't enough! -- On the Effectiveness of Debiasing MLMs and their Social Biases in Downstream Tasks","created":"2022-10-06","cited_by_count":48,"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":"2301.10884","title":"Break It Down: Evidence for Structural Compositionality in Neural Networks","created":"2023-01-26","cited_by_count":44,"area":null},{"id":"2207.04153","title":"Probing Classifiers are Unreliable for Concept Removal and Detection","created":"2022-07-08","cited_by_count":41,"area":null},{"id":"2502.07218","title":"LLM Unlearning via Neural Activation Redirection","created":"2025-02-11","cited_by_count":41,"area":"safety"},{"id":"2105.07109","title":"The Low-Dimensional Linear Geometry of Contextualized Word Representations","created":"2021-05-15","cited_by_count":40,"area":null},{"id":"2404.02062","title":"Digital Forgetting in Large Language Models: A Survey of Unlearning Methods","created":"2024-04-02","cited_by_count":40,"area":"safety"},{"id":"2504.14492","title":"FairSteer: Inference Time Debiasing for LLMs with Dynamic Activation Steering","created":"2025-04-20","cited_by_count":40,"area":"mechinterp"},{"id":"2507.16795","title":"Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning","created":"2025-07-22","cited_by_count":39,"area":null},{"id":"2305.12090","title":"UP5: Unbiased Foundation Model for Fairness-aware Recommendation","created":"2023-05-20","cited_by_count":38,"area":null},{"id":"2212.08094","title":"Joint processing of linguistic properties in brains and language models","created":"2022-12-15","cited_by_count":37,"area":null},{"id":"2507.08802","title":"The Non-Linear Representation Dilemma: Is Causal Abstraction Enough for Mechanistic Interpretability?","created":"2025-07-11","cited_by_count":37,"area":"mechinterp"},{"id":"2406.09413","title":"Interpreting the Weight Space of Customized Diffusion Models","created":"2024-06-13","cited_by_count":36,"area":"diffusion"},{"id":"2104.05196","title":"StylePTB: A Compositional Benchmark for Fine-grained Controllable Text Style Transfer","created":"2021-04-12","cited_by_count":34,"area":null},{"id":"2109.08253","title":"Balancing out Bias: Achieving Fairness Through Balanced Training","created":"2021-09-16","cited_by_count":34,"area":null},{"id":"2109.10645","title":"Contrastive Learning for Fair Representations","created":"2021-09-22","cited_by_count":34,"area":null},{"id":"2203.07893","title":"Gold Doesn't Always Glitter: Spectral Removal of Linear and Nonlinear Guarded Attribute Information","created":"2022-03-15","cited_by_count":34,"area":null},{"id":"2410.04332","title":"Gradient Routing: Masking Gradients to Localize Computation in Neural Networks","created":"2024-10-06","cited_by_count":34,"area":null},{"id":"2210.14975","title":"MABEL: Attenuating Gender Bias using Textual Entailment Data","created":"2022-10-26","cited_by_count":31,"area":null},{"id":"2407.00600","title":"GenderBias-\\emph{VL}: Benchmarking Gender Bias in Vision Language Models via Counterfactual Probing","created":"2024-06-30","cited_by_count":31,"area":"mllm"},{"id":"2109.10441","title":"Evaluating Debiasing Techniques for Intersectional Biases","created":"2021-09-21","cited_by_count":30,"area":null},{"id":"2209.14279","title":"Causal Proxy Models for Concept-Based Model Explanations","created":"2022-09-28","cited_by_count":29,"area":null},{"id":"2205.15301","title":"Can Transformer be Too Compositional? Analysing Idiom Processing in Neural Machine Translation","created":"2022-05-30","cited_by_count":28,"area":null},{"id":"2210.14011","title":"Are All Spurious Features in Natural Language Alike? An Analysis through a Causal Lens","created":"2022-10-25","cited_by_count":28,"area":null},{"id":"2306.04597","title":"Language Models Get a Gender Makeover: Mitigating Gender Bias with Few-Shot Data Interventions","created":"2023-06-07","cited_by_count":28,"area":null},{"id":"2306.08158","title":"Sociodemographic Bias in Language Models: A Survey and Forward Path","created":"2023-06-13","cited_by_count":28,"area":null},{"id":"2310.11958","title":"Emptying the Ocean with a Spoon: Should We Edit Models?","created":"2023-10-18","cited_by_count":27,"area":null},{"id":"2406.11107","title":"Exploring Safety-Utility Trade-Offs in Personalized Language Models","created":"2024-06-17","cited_by_count":27,"area":null},{"id":"2406.01285","title":"Large Language Models as Recommender Systems: A Study of Popularity Bias","created":"2024-06-03","cited_by_count":26,"area":null},{"id":"2409.08087","title":"Securing Large Language Models: Addressing Bias, Misinformation, and Prompt Attacks","created":"2024-09-12","cited_by_count":26,"area":"safety"},{"id":"2302.06321","title":"Parameter-efficient Modularised Bias Mitigation via AdapterFusion","created":"2023-02-13","cited_by_count":25,"area":null},{"id":"2305.10204","title":"Shielded Representations: Protecting Sensitive Attributes Through Iterative Gradient-Based Projection","created":"2023-05-17","cited_by_count":25,"area":null},{"id":"2307.10514","title":"Building Socio-culturally Inclusive Stereotype Resources with Community Engagement","created":"2023-07-20","cited_by_count":25,"area":null},{"id":"2307.15054","title":"A Geometric Notion of Causal Probing","created":"2023-07-27","cited_by_count":25,"area":null},{"id":"2402.17954","title":"Twists, Humps, and Pebbles: Multilingual Speech Recognition Models Exhibit Gender Performance Gaps","created":"2024-02-28","cited_by_count":25,"area":null},{"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":"2101.09525","title":"Dictionary-based Debiasing of Pre-trained Word Embeddings","created":"2021-01-23","cited_by_count":24,"area":null},{"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":"2208.10244","title":"Unit Testing for Concepts in Neural Networks","created":"2022-07-28","cited_by_count":23,"area":null},{"id":"2302.13319","title":"Efficient fair PCA for fair representation learning","created":"2023-02-26","cited_by_count":23,"area":null},{"id":"2310.16484","title":"Subspace Chronicles: How Linguistic Information Emerges, Shifts and Interacts during Language Model Training","created":"2023-10-25","cited_by_count":23,"area":null},{"id":"2402.06900","title":"Can LLMs Recognize Toxicity? A Structured Investigation Framework and Toxicity Metric","created":"2024-02-10","cited_by_count":23,"area":null},{"id":"2406.05392","title":"Deconstructing The Ethics of Large Language Models from Long-standing Issues to New-emerging Dilemmas: A Survey","created":"2024-06-08","cited_by_count":23,"area":null},{"id":"2407.04690","title":"Missed Causes and Ambiguous Effects: Counterfactuals Pose Challenges for Interpreting Neural Networks","created":"2024-07-05","cited_by_count":23,"area":null},{"id":"2504.12459","title":"On Linear Representations and Pretraining Data Frequency in Language Models","created":"2025-04-16","cited_by_count":23,"area":null},{"id":"2205.02393","title":"Optimising Equal Opportunity Fairness in Model Training","created":"2022-05-05","cited_by_count":22,"area":null},{"id":"2210.07269","title":"SODAPOP: Open-Ended Discovery of Social Biases in Social Commonsense Reasoning Models","created":"2022-10-13","cited_by_count":22,"area":"lr"},{"id":"2212.10563","title":"BLIND: Bias Removal With No Demographics","created":"2022-12-20","cited_by_count":22,"area":null},{"id":"2310.12611","title":"Identifying and Adapting Transformer-Components Responsible for Gender Bias in an English Language Model","created":"2023-10-19","cited_by_count":22,"area":null},{"id":"2602.07050","title":"Interpreting Physics in Video World Models","created":"2026-02-04","cited_by_count":22,"area":"wm"},{"id":"2307.01595","title":"Prompt Tuning Pushes Farther, Contrastive Learning Pulls Closer: A Two-Stage Approach to Mitigate Social Biases","created":"2023-07-04","cited_by_count":21,"area":null},{"id":"2202.06618","title":"A Differential Entropy Estimator for Training Neural Networks","created":"2022-02-14","cited_by_count":20,"area":null},{"id":"2307.03360","title":"Evaluating Biased Attitude Associations of Language Models in an Intersectional Context","created":"2023-07-07","cited_by_count":20,"area":null}],"next_cursor":"100"}