{"total":192,"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":"2406.04313","title":"Improving Alignment and Robustness with Circuit Breakers","created":"2024-06-06","cited_by_count":322,"area":"safety"},{"id":"2406.18510","title":"WildTeaming at Scale: From In-the-Wild Jailbreaks to (Adversarially) Safer Language Models","created":"2024-06-26","cited_by_count":223,"area":"safety"},{"id":"2402.16822","title":"Rainbow Teaming: Open-Ended Generation of Diverse Adversarial Prompts","created":"2024-02-26","cited_by_count":205,"area":"safety"},{"id":"2307.16888","title":"Backdooring Instruction-Tuned Large Language Models with Virtual Prompt Injection","created":"2023-07-31","cited_by_count":157,"area":"safety"},{"id":"2402.15159","title":"Machine Unlearning of Pre-trained Large Language Models","created":"2024-02-23","cited_by_count":119,"area":"safety"},{"id":"2404.01099","title":"What is in Your Safe Data? Identifying Benign Data that Breaks Safety","created":"2024-04-01","cited_by_count":118,"area":"safety"},{"id":"2308.00113","title":"Three Bricks to Consolidate Watermarks for Large Language Models","created":"2023-07-26","cited_by_count":84,"area":"safety"},{"id":"2506.12618","title":"OpenUnlearning: Accelerating LLM Unlearning via Unified Benchmarking of Methods and Metrics","created":"2025-06-14","cited_by_count":83,"area":"safety"},{"id":"2502.05374","title":"Towards LLM Unlearning Resilient to Relearning Attacks: A Sharpness-Aware Minimization Perspective and Beyond","created":"2025-02-07","cited_by_count":79,"area":"safety"},{"id":"2402.09179","title":"Instruction Backdoor Attacks Against Customized LLMs","created":"2024-02-14","cited_by_count":62,"area":"safety"},{"id":"2402.14904","title":"Watermarking Makes Language Models Radioactive","created":"2024-02-22","cited_by_count":58,"area":"safety"},{"id":"2406.17092","title":"BEEAR: Embedding-based Adversarial Removal of Safety Backdoors in Instruction-tuned Language Models","created":"2024-06-24","cited_by_count":58,"area":"safety"},{"id":"2410.08109","title":"A Closer Look at Machine Unlearning for Large Language Models","created":"2024-10-10","cited_by_count":56,"area":"safety"},{"id":"2405.18540","title":"Learning diverse attacks on large language models for robust red-teaming and safety tuning","created":"2024-05-28","cited_by_count":54,"area":"safety"},{"id":"2408.12798","title":"BackdoorLLM: A Comprehensive Benchmark for Backdoor Attacks and Defenses on Large Language Models","created":"2024-08-23","cited_by_count":50,"area":"safety"},{"id":"2506.07022","title":"AlphaSteer: Learning Refusal Steering with Principled Null-Space Constraint","created":"2025-06-08","cited_by_count":49,"area":"safety"},{"id":"2505.16831","title":"Unlearning Isn't Deletion: Investigating Reversibility of Machine Unlearning in LLMs","created":"2025-05-22","cited_by_count":42,"area":"safety"},{"id":"2405.09055","title":"A safety realignment framework via subspace-oriented model fusion for large language models","created":"2024-05-15","cited_by_count":41,"area":"safety"},{"id":"2403.01251","title":"Accelerating Greedy Coordinate Gradient and General Prompt Optimization via Probe Sampling","created":"2024-03-02","cited_by_count":37,"area":"safety"},{"id":"2308.08090","title":"Separate the Wheat from the Chaff: Model Deficiency Unlearning via Parameter-Efficient Module Operation","created":"2023-08-16","cited_by_count":33,"area":"safety"},{"id":"2504.01550","title":"Representation Bending for Large Language Model Safety","created":"2025-04-02","cited_by_count":30,"area":"safety"},{"id":"2406.05948","title":"Chain-of-Scrutiny: Detecting Backdoor Attacks for Large Language Models","created":"2024-06-10","cited_by_count":28,"area":"safety"},{"id":"2504.10694","title":"The Jailbreak Tax: How Useful are Your Jailbreak Outputs?","created":"2025-04-14","cited_by_count":28,"area":"safety"},{"id":"2401.10862","title":"Pruning for Protection: Increasing Jailbreak Resistance in Aligned LLMs Without Fine-Tuning","created":"2024-01-19","cited_by_count":26,"area":"safety"},{"id":"2409.07314","title":"MEDIC: Comprehensive Evaluation of Leading Indicators for LLM Safety and Utility in Clinical Applications","created":"2024-09-11","cited_by_count":25,"area":"safety"},{"id":"2505.11541","title":"MorphMark: Flexible Adaptive Watermarking for Large Language Models","created":"2025-05-14","cited_by_count":24,"area":"safety"},{"id":"2407.20224","title":"Can Editing LLMs Inject Harm?","created":"2024-07-29","cited_by_count":22,"area":"safety"},{"id":"2410.14827","title":"Enhancing Prompt Injection Attacks to LLMs via Poisoning Alignment","created":"2024-10-18","cited_by_count":21,"area":"safety"},{"id":"2404.05188","title":"Have You Merged My Model? On The Robustness of Large Language Model IP Protection Methods Against Model Merging","created":"2024-04-08","cited_by_count":19,"area":"safety"},{"id":"2407.00869","title":"Large Language Models Are Involuntary Truth-Tellers: Exploiting Fallacy Failure for Jailbreak Attacks","created":"2024-07-01","cited_by_count":19,"area":"safety"},{"id":"2406.11780","title":"Split, Unlearn, Merge: Leveraging Data Attributes for More Effective Unlearning in LLMs","created":"2024-06-17","cited_by_count":18,"area":"safety"},{"id":"2505.24428","title":"Model Unlearning via Sparse Autoencoder Subspace Guided Projections","created":"2025-05-30","cited_by_count":18,"area":"safety"},{"id":"2506.01114","title":"Reconsidering LLM Uncertainty Estimation Methods in the Wild","created":"2025-06-01","cited_by_count":17,"area":"safety"},{"id":"2507.09709","title":"Large Language Models Encode Semantics and Alignment in Linearly Separable Representations","created":"2025-07-13","cited_by_count":17,"area":"safety"},{"id":"2510.21189","title":"Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency","created":"2025-10-24","cited_by_count":17,"area":"safety"},{"id":"2502.11441","title":"Which Retain Set Matters for LLM Unlearning? A Case Study on Entity Unlearning","created":"2025-02-17","cited_by_count":16,"area":"safety"},{"id":"2502.13141","title":"UniGuardian: A Unified Defense for Detecting Prompt Injection, Backdoor Attacks and Adversarial Attacks in Large Language Models","created":"2025-02-18","cited_by_count":16,"area":"safety"},{"id":"2506.07171","title":"RULE: Reinforcement UnLEarning Achieves Forget-Retain Pareto Optimality","created":"2025-06-08","cited_by_count":16,"area":"safety"},{"id":"2411.16736","title":"ChemSafetyBench: Benchmarking LLM Safety on Chemistry Domain","created":"2024-11-23","cited_by_count":15,"area":"safety"},{"id":"2406.09325","title":"REVS: Unlearning Sensitive Information in Language Models via Rank Editing in the Vocabulary Space","created":"2024-06-13","cited_by_count":13,"area":"safety"},{"id":"2406.13748","title":"Learn and Unlearn: Addressing Misinformation in Multilingual LLMs","created":"2024-06-19","cited_by_count":12,"area":"safety"},{"id":"2410.06467","title":"WAPITI: A Watermark for Finetuned Open-Source LLMs","created":"2024-10-09","cited_by_count":12,"area":"safety"},{"id":"2502.16903","title":"GuidedBench: Measuring and Mitigating the Evaluation Discrepancies of In-the-wild LLM Jailbreak Methods","created":"2025-02-24","cited_by_count":12,"area":"safety"},{"id":"2504.10185","title":"LLM Unlearning Reveals a Stronger-Than-Expected Coreset Effect in Current Benchmarks","created":"2025-04-14","cited_by_count":12,"area":"safety"},{"id":"2506.15606","title":"LoX: Low-Rank Extrapolation Robustifies LLM Safety Against Fine-tuning","created":"2025-06-18","cited_by_count":12,"area":"safety"},{"id":"2509.24296","title":"DiffuGuard: How Intrinsic Safety is Lost and Found in Diffusion Large Language Models","created":"2025-09-29","cited_by_count":12,"area":"safety"},{"id":"2407.01376","title":"Badllama 3: removing safety finetuning from Llama 3 in minutes","created":"2024-07-01","cited_by_count":11,"area":"safety"},{"id":"2502.18511","title":"ELBA-Bench: An Efficient Learning Backdoor Attacks Benchmark for Large Language Models","created":"2025-02-22","cited_by_count":11,"area":"safety"},{"id":"2504.07113","title":"How Robust Are Router-LLMs? Analysis of the Fragility of LLM Routing Capabilities","created":"2025-03-20","cited_by_count":11,"area":"safety"},{"id":"2505.16559","title":"CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning","created":"2025-05-22","cited_by_count":11,"area":"safety"},{"id":"2508.12398","title":"Where to Start Alignment? Diffusion Large Language Model May Demand a Distinct Position","created":"2025-08-17","cited_by_count":11,"area":"safety"},{"id":"2502.11647","title":"DELMAN: Dynamic Defense Against Large Language Model Jailbreaking with Model Editing","created":"2025-02-17","cited_by_count":10,"area":"safety"},{"id":"2503.04693","title":"UIPE: Enhancing LLM Unlearning by Removing Knowledge Related to Forgetting Targets","created":"2025-03-06","cited_by_count":10,"area":"safety"},{"id":"2505.12186","title":"Self-Destructive Language Model","created":"2025-05-18","cited_by_count":10,"area":"safety"},{"id":"2501.19202","title":"Improving LLM Unlearning Robustness via Random Perturbations","created":"2025-01-31","cited_by_count":9,"area":"safety"},{"id":"2503.21730","title":"Effective Skill Unlearning through Intervention and Abstention","created":"2025-03-27","cited_by_count":9,"area":"safety"},{"id":"2505.16723","title":"LLM Fingerprinting via Semantically Conditioned Watermarks","created":"2025-05-22","cited_by_count":9,"area":"safety"},{"id":"2507.01321","title":"ICLShield: Exploring and Mitigating In-Context Learning Backdoor Attacks","created":"2025-07-02","cited_by_count":9,"area":"safety"},{"id":"2504.19674","title":"SAGE: A Generic Framework for LLM Safety Evaluation","created":"2025-04-28","cited_by_count":8,"area":"safety"},{"id":"2505.12060","title":"Why Not Act on What You Know? Unleashing Safety Potential of LLMs via Self-Aware Guard Enhancement","created":"2025-05-17","cited_by_count":8,"area":"safety"},{"id":"2505.16530","title":"DuFFin: A Dual-Level Fingerprinting Framework for LLMs IP Protection","created":"2025-05-22","cited_by_count":8,"area":"safety"},{"id":"2510.07835","title":"MetaDefense: Defending Finetuning-based Jailbreak Attack Before and During Generation","created":"2025-10-09","cited_by_count":8,"area":"safety"},{"id":"2505.15753","title":"Scalable Defense against In-the-wild Jailbreaking Attacks with Safety Context Retrieval","created":"2025-05-21","cited_by_count":7,"area":"safety"},{"id":"2505.23561","title":"Merge Hijacking: Backdoor Attacks to Model Merging of Large Language Models","created":"2025-05-29","cited_by_count":7,"area":"safety"},{"id":"2506.14682","title":"AIRTBench: Measuring Autonomous AI Red Teaming Capabilities in Language Models","created":"2025-06-17","cited_by_count":7,"area":"safety"},{"id":"2512.14741","title":"Persistent Backdoor Attacks under Continual Fine-Tuning of LLMs","created":"2025-12-12","cited_by_count":7,"area":"safety"},{"id":"2601.07087","title":"When Should We Introduce Safety Interventions During Pretraining?","created":"2026-01-11","cited_by_count":7,"area":"safety"},{"id":"2601.10141","title":"Understanding and Preserving Safety in Fine-Tuned LLMs","created":"2026-01-15","cited_by_count":7,"area":"safety"},{"id":"2603.13461","title":"Purifying Generative LLMs from Backdoors without Prior Knowledge or Clean Reference","created":"2026-03-13","cited_by_count":7,"area":"safety"},{"id":"2402.13517","title":"Round Trip Translation Defence against Large Language Model Jailbreaking Attacks","created":"2024-02-21","cited_by_count":6,"area":"safety"},{"id":"2504.14738","title":"PROMPTEVALS: A Dataset of Assertions and Guardrails for Custom Production Large Language Model Pipelines","created":"2025-04-20","cited_by_count":6,"area":"safety"},{"id":"2510.16629","title":"On the Impossibility of Retrain Equivalence in Machine Unlearning","created":"2025-10-18","cited_by_count":6,"area":"safety"},{"id":"2511.07863","title":"WaterMod: Modular Token-Rank Partitioning for Probability-Balanced LLM Watermarking","created":"2025-11-11","cited_by_count":6,"area":"safety"},{"id":"2512.16602","title":"Refusal Steering: Fine-grained Control over LLM Refusal Behaviour for Sensitive Topics","created":"2025-12-18","cited_by_count":6,"area":"safety"},{"id":"2602.15143","title":"Protecting Language Models Against Unauthorized Distillation through Trace Rewriting","created":"2026-02-16","cited_by_count":6,"area":"safety"},{"id":"2603.00498","title":"Antibody: Strengthening Defense Against Harmful Fine-Tuning for Large Language Models via Attenuating Harmful Gradient Influence","created":"2026-02-28","cited_by_count":6,"area":"safety"},{"id":"2607.19829","title":"DARWIN: Evolving Jailbreak Adversary and Guardrail for LLM Safety Evaluation and Protection","created":"2026-07-22","cited_by_count":6,"area":"safety"},{"id":"2402.05723","title":"In-Context Learning Can Re-learn Forbidden Tasks","created":"2024-02-08","cited_by_count":5,"area":"safety"},{"id":"2410.10150","title":"Jailbreak Instruction-Tuned LLMs via end-of-sentence MLP Re-weighting","created":"2024-10-14","cited_by_count":5,"area":"safety"},{"id":"2411.19335","title":"PEFT-as-an-Attack! Jailbreaking Language Models during Federated Parameter-Efficient Fine-Tuning","created":"2024-11-28","cited_by_count":5,"area":"safety"},{"id":"2503.11185","title":"Bleeding Pathways: Vanishing Discriminability in LLM Hidden States Fuels Jailbreak Attacks","created":"2025-03-14","cited_by_count":5,"area":"safety"},{"id":"2505.15674","title":"UniErase: Towards Balanced and Precise Unlearning in Language Models","created":"2025-05-21","cited_by_count":5,"area":"safety"},{"id":"2505.16567","title":"Watch your steps: Dormant Adversarial Behaviors that Activate upon LLM Finetuning","created":"2025-05-22","cited_by_count":5,"area":"safety"},{"id":"2508.06595","title":"LLM Unlearning Without an Expert Curated Dataset","created":"2025-08-08","cited_by_count":5,"area":"safety"},{"id":"2509.22263","title":"Erase or Hide? Suppressing Spurious Unlearning Neurons for Robust Unlearning","created":"2025-09-26","cited_by_count":5,"area":"safety"},{"id":"2510.02962","title":"Leave No TRACE: Black-box Detection of Copyrighted Dataset Usage in Large Language Models via Watermarking","created":"2025-10-03","cited_by_count":5,"area":"safety"},{"id":"2511.06890","title":"EduGuardBench: A Holistic Benchmark for Evaluating the Pedagogical Fidelity and Adversarial Safety of LLMs as Simulated Teachers","created":"2025-11-10","cited_by_count":5,"area":"safety"},{"id":"2601.08089","title":"Q-realign: Piggybacking Realignment on Quantization for Safe and Efficient LLM Deployment","created":"2026-01-13","cited_by_count":5,"area":"safety"},{"id":"2602.06754","title":"A Unified Framework for LLM Watermarks","created":"2026-02-06","cited_by_count":5,"area":"safety"},{"id":"2604.15022","title":"Route to Rome Attack: Directing LLM Routers to Expensive Models via Adversarial Suffix Optimization","created":"2026-04-16","cited_by_count":5,"area":"safety"},{"id":"2412.03107","title":"CredID: Credible Multi-Bit Watermark for Large Language Models Identification","created":"2024-12-04","cited_by_count":4,"area":"safety"},{"id":"2502.16174","title":"Efficient LLM Moderation with Multi-Layer Latent Prototypes","created":"2025-02-22","cited_by_count":4,"area":"safety"},{"id":"2503.01926","title":"Unnatural Languages Are Not Bugs but Features for LLMs","created":"2025-03-02","cited_by_count":4,"area":"safety"},{"id":"2505.13500","title":"Noise Injection Systemically Degrades Large Language Model Safety Guardrails","created":"2025-05-16","cited_by_count":4,"area":"safety"},{"id":"2505.15209","title":"DUSK: Do Not Unlearn Shared Knowledge","created":"2025-05-21","cited_by_count":4,"area":"safety"},{"id":"2507.19894","title":"Generative Model Unlearning: A Survey through Target Events, Unlearning Operators, and Evaluation Protocols","created":"2025-07-26","cited_by_count":4,"area":"safety"},{"id":"2508.12072","title":"Mitigating Jailbreaks with Intent-Aware LLMs","created":"2025-08-16","cited_by_count":4,"area":"safety"},{"id":"2508.13214","title":"Too Easily Fooled? Prompt Injection Breaks LLMs on Frustratingly Simple Multiple-Choice Questions","created":"2025-08-16","cited_by_count":4,"area":"safety"},{"id":"2509.08000","title":"AntiDote: Bi-level Adversarial Training for Tamper-Resistant LLMs","created":"2025-09-06","cited_by_count":4,"area":"safety"}],"next_cursor":"100"}