{"total":79,"rows":[{"id":"2408.15221","title":"LLM Defenses Are Not Robust to Multi-Turn Human Jailbreaks Yet","created":"2024-08-27","cited_by_count":151,"area":"safety"},{"id":"2410.18927","title":"SafeBench: A Safety Evaluation Framework for Multimodal Large Language Models","created":"2024-10-24","cited_by_count":69,"area":"safety"},{"id":"2504.17704","title":"Safety in Large Reasoning Models: A Survey","created":"2025-04-24","cited_by_count":64,"area":"lr"},{"id":"2501.18841","title":"Trading Inference-Time Compute for Adversarial Robustness","created":"2025-01-31","cited_by_count":61,"area":"lr"},{"id":"2503.15754","title":"AutoRedTeamer: Autonomous Red Teaming with Lifelong Attack Integration","created":"2025-03-20","cited_by_count":43,"area":"safety"},{"id":"2502.12970","title":"Reasoning-to-Defend: Safety-Aware Reasoning Can Defend Large Language Models from Jailbreaking","created":"2025-02-18","cited_by_count":34,"area":"lr"},{"id":"2504.01550","title":"Representation Bending for Large Language Model Safety","created":"2025-04-02","cited_by_count":30,"area":"safety"},{"id":"2502.11355","title":"Nuclear Deployed: Analyzing Catastrophic Risks in Decision-making of Autonomous LLM Agents","created":"2025-02-17","cited_by_count":27,"area":"lr"},{"id":"2410.02298","title":"Jailbreak Antidote: Runtime Safety-Utility Balance via Sparse Representation Adjustment in Large Language Models","created":"2024-10-03","cited_by_count":23,"area":"safety"},{"id":"2410.01294","title":"Endless Jailbreaks with Bijection Learning","created":"2024-10-02","cited_by_count":19,"area":"safety"},{"id":"2503.02574","title":"LLM-Safety Evaluations Lack Robustness","created":"2025-03-04","cited_by_count":18,"area":"alignment"},{"id":"2408.03837","title":"WalledEval: A Comprehensive Safety Evaluation Toolkit for Large Language Models","created":"2024-08-07","cited_by_count":17,"area":null},{"id":"2603.12277","title":"Prompt Injection as Role Confusion","created":"2026-02-22","cited_by_count":17,"area":"safety"},{"id":"2505.13527","title":"Logic Jailbreak: Efficiently Unlocking LLM Safety Restrictions Through Formal Logical Expression","created":"2025-05-18","cited_by_count":16,"area":"alignment"},{"id":"2502.04040","title":"Safety Reasoning with Guidelines","created":"2025-02-06","cited_by_count":15,"area":"lr"},{"id":"2502.04322","title":"Speak Easy: Eliciting Harmful Jailbreaks from LLMs with Simple Interactions","created":"2025-02-06","cited_by_count":15,"area":"alignment"},{"id":"2505.20259","title":"Lifelong Safety Alignment for Language Models","created":"2025-05-26","cited_by_count":15,"area":"alignment"},{"id":"2507.14987","title":"AlphaAlign: Incentivizing Safety Alignment with Extremely Simplified Reinforcement Learning","created":"2025-07-20","cited_by_count":15,"area":"alignment"},{"id":"2410.03857","title":"You Know What I'm Saying: Jailbreak Attack via Implicit Reference","created":"2024-10-04","cited_by_count":14,"area":"safety"},{"id":"2503.08195","title":"Dialogue Injection Attack: Jailbreaking LLMs through Context Manipulation","created":"2025-03-11","cited_by_count":14,"area":"safety"},{"id":"2512.05648","title":"Beyond Data Filtering: Knowledge Localization for Capability Removal in LLMs","created":"2025-12-05","cited_by_count":11,"area":"safety"},{"id":"2410.22143","title":"AmpleGCG-Plus: A Strong Generative Model of Adversarial Suffixes to Jailbreak LLMs with Higher Success Rates in Fewer Attempts","created":"2024-10-29","cited_by_count":10,"area":"safety"},{"id":"2502.11910","title":"Adversarial Alignment for LLMs Requires Simpler, Reproducible, and More Measurable Objectives","created":"2025-02-17","cited_by_count":8,"area":null},{"id":"2501.18626","title":"The TIP of the Iceberg: Revealing a Hidden Class of Task-in-Prompt Adversarial Attacks on LLMs","created":"2025-01-27","cited_by_count":7,"area":"safety"},{"id":"2512.05485","title":"TeleAI-Safety: A comprehensive LLM jailbreaking benchmark towards attacks, defenses, and evaluations","created":"2025-12-05","cited_by_count":7,"area":"safety"},{"id":"2409.03274","title":"Recent Advances in Attack and Defense Approaches of Large Language Models","created":"2024-09-05","cited_by_count":6,"area":null},{"id":"2410.13691","title":"Jailbreaking LLM-Controlled Robots","created":"2024-10-17","cited_by_count":6,"area":"safety"},{"id":"2411.02785","title":"Stochastic Monkeys at Play: Random Augmentations Cheaply Break LLM Safety Alignment","created":"2024-11-05","cited_by_count":5,"area":"alignment"},{"id":"2510.15476","title":"SoK: Systematizing LLM Prompt Security: Taxonomies, Datasets, and Unified Evaluation of Attacks and Defenses","created":"2025-10-17","cited_by_count":5,"area":"safety"},{"id":"2410.04190","title":"Harnessing Task Overload for Scalable Jailbreak Attacks on Large Language Models","created":"2024-10-05","cited_by_count":4,"area":"safety"},{"id":"2411.01084","title":"Plentiful Jailbreaks with String Compositions","created":"2024-11-01","cited_by_count":4,"area":"safety"},{"id":"2502.13603","title":"Efficient Safety Retrofitting Against Jailbreaking for LLMs","created":"2025-02-19","cited_by_count":4,"area":"alignment"},{"id":"2505.13862","title":"PandaGuard: Systematic Evaluation of LLM Safety against Jailbreaking Attacks","created":"2025-05-20","cited_by_count":4,"area":"alignment"},{"id":"2505.24672","title":"TRIDENT: Enhancing Large Language Model Safety with Tri-Dimensional Diversified Red-Teaming Data Synthesis","created":"2025-05-30","cited_by_count":4,"area":"synth"},{"id":"2509.25271","title":"RADAR: A Risk-Aware Dynamic Multi-Agent Framework for LLM Safety Evaluation via Role-Specialized Collaboration","created":"2025-09-28","cited_by_count":4,"area":"agents"},{"id":"2510.01088","title":"Safety Instincts: LLMs Learn to Trust Their Internal Compass for Self-Defense","created":"2025-10-01","cited_by_count":4,"area":"safety"},{"id":"2509.19360","title":"Semantic Representation Attack against Aligned Large Language Models","created":"2025-09-18","cited_by_count":3,"area":"alignment"},{"id":"2510.21190","title":"The Trojan Example: Jailbreaking LLMs through Template Filling and Unsafety Reasoning","created":"2025-10-24","cited_by_count":3,"area":"alignment"},{"id":"2511.01187","title":"Surfacing Subtle Stereotypes: A Multilingual, Debate-Oriented Evaluation of Modern LLMs","created":"2025-11-03","cited_by_count":3,"area":null},{"id":"2602.03608","title":"Controlling Output Rankings in Generative Engines for LLM-based Search","created":"2026-02-03","cited_by_count":3,"area":null},{"id":"2606.06533","title":"Position: Don't Just \"Fix it in Post\": A Science of AI Must Study Training Dynamics","created":"2026-06-03","cited_by_count":3,"area":null},{"id":"2504.05050","title":"Revealing the Intrinsic Ethical Vulnerability of Aligned Large Language Models","created":"2025-04-07","cited_by_count":2,"area":"alignment"},{"id":"2504.09712","title":"The Structural Safety Generalization Problem","created":"2025-04-13","cited_by_count":2,"area":"safety"},{"id":"2506.12274","title":"InfoFlood: Jailbreaking Large Language Models with Information Overload","created":"2025-06-13","cited_by_count":2,"area":"safety"},{"id":"2511.06396","title":"Efficient LLM Safety Evaluation through Multi-Agent Debate","created":"2025-11-09","cited_by_count":2,"area":"agents"},{"id":"2602.05164","title":"Position: Capability Control Should be a Separate Goal From Alignment","created":"2026-02-05","cited_by_count":2,"area":null},{"id":"2602.24009","title":"Jailbreak Foundry: From Papers to Runnable Attacks for Reproducible Benchmarking","created":"2026-02-27","cited_by_count":2,"area":"safety"},{"id":"2506.07031","title":"HauntAttack: When Attack Follows Reasoning as a Shadow","created":"2025-06-08","cited_by_count":1,"area":"lr"},{"id":"2508.10404","title":"Layer-Wise Perturbations via Sparse Autoencoders for Adversarial Text Generation","created":"2025-08-14","cited_by_count":1,"area":"safety"},{"id":"2508.15526","title":"SafetyFlow: An Agent-Flow System for Automated LLM Safety Benchmarking","created":"2025-08-21","cited_by_count":1,"area":"agents"},{"id":"2509.22732","title":"Bidirectional Intention Inference Enhances LLMs' Defense Against Multi-Turn Jailbreak Attacks","created":"2025-09-25","cited_by_count":1,"area":"alignment"},{"id":"2509.25843","title":"ASGuard: Activation-Scaling Guard to Mitigate Targeted Jailbreaking Attack","created":"2025-09-30","cited_by_count":1,"area":"safety"},{"id":"2601.19507","title":"Automated Safety Benchmarking: A Multi-agent Pipeline for LVLMs","created":"2026-01-27","cited_by_count":1,"area":"mllm"},{"id":"2606.02644","title":"A New Framework for Cybersecurity Refusals in AI Agents","created":"2026-05-31","cited_by_count":1,"area":"agents"},{"id":"2607.25379","title":"Cyber-Capable AI Agents: Vulnerabilities, Evaluation Containment, and Defensive Response","created":"2026-07-28","cited_by_count":1,"area":"agents"},{"id":"2503.00224","title":"\\`A la recherche du sens perdu: your favourite LLM might have more to say than you can understand","created":"2025-02-28","cited_by_count":0,"area":null},{"id":"2506.06391","title":"From Rogue to Safe AI: The Role of Explicit Refusals in Aligning LLMs with International Humanitarian Law","created":"2025-06-05","cited_by_count":0,"area":null},{"id":"2508.06296","title":"LLM Robustness Leaderboard v1 --Technical report","created":"2025-08-08","cited_by_count":0,"area":"safety"},{"id":"2508.13246","title":"Involuntary Jailbreak: On Self-Prompting Attacks","created":"2025-08-18","cited_by_count":0,"area":"safety"},{"id":"2510.01259","title":"In AI Sweet Harmony: Sociopragmatic Guardrail Bypasses and Evaluation-Awareness in OpenAI gpt-oss-20b","created":"2025-09-25","cited_by_count":0,"area":"safety"},{"id":"2510.25941","title":"RECAP: Reproducing Copyrighted Data from LLMs Training with an Agentic Pipeline","created":"2025-10-29","cited_by_count":0,"area":"agents"},{"id":"2511.22044","title":"Distillability of LLM Security Logic: Predicting Attack Success Rate of Outline Filling Attack via Ranking Regression","created":"2025-11-27","cited_by_count":0,"area":"safety"},{"id":"2512.24556","title":"Safe in the Future, Dangerous in the Past: Dissecting Temporal and Linguistic Vulnerabilities in LLMs","created":"2025-12-31","cited_by_count":0,"area":"alignment"},{"id":"2601.12652","title":"Ethical Risks in Deploying Large Language Models: An Evaluation of Medical Ethics Jailbreaking","created":"2026-01-19","cited_by_count":0,"area":"safety"},{"id":"2602.15238","title":"Closing the Distribution Gap in Adversarial Training for LLMs","created":"2026-02-16","cited_by_count":0,"area":null},{"id":"2605.08898","title":"LLM-Agnostic Semantic Representation Attack","created":"2026-05-09","cited_by_count":0,"area":"safety"},{"id":"2605.27545","title":"PAST2HARM: A Simple Adaptive Past Tense Attack for Jailbreaking Multimodal AI","created":"2026-05-26","cited_by_count":0,"area":"safety"},{"id":"2606.24388","title":"PHANTOM: A Large-Scale Dataset of Multimodal Adversarial Attacks for Vision-Language Models","created":"2026-06-23","cited_by_count":0,"area":"mllm"},{"id":"2607.02961","title":"Overloading Large Vision-Language Models for Jailbreaking","created":"2026-07-03","cited_by_count":0,"area":"mllm"},{"id":"2607.04645","title":"Retroactive Chain-of-Thought (RetroCoT): Forensic Reconstruction Prompts as a Safety Diagnostic Across Model Generations","created":"2026-07-06","cited_by_count":0,"area":"lr"},{"id":"2607.17152","title":"How Jailbreak Attacks Inform Safety Alignment: A Defender-Centric, Shapley-Based Evaluation of Jailbreak Contributions","created":"2026-07-19","cited_by_count":0,"area":"safety"},{"id":"2607.21735","title":"What AI Red-Team Evaluations Can and Cannot Prove","created":"2026-07-23","cited_by_count":0,"area":"safety"},{"id":"2608.05409","title":"Mood Matters: How Syntactic Sensitivity Undermines Safety Alignment","created":"2026-08-05","cited_by_count":0,"area":"alignment"},{"id":"2609.02707","title":"Door-in-the-Face Requests and Refusal Behaviour in Large Language Models","created":"2026-09-02","cited_by_count":0,"area":null},{"id":"2609.04022","title":"Representational alignment yields generalizable safety in language models","created":"2026-09-03","cited_by_count":0,"area":"alignment"},{"id":"2609.05797","title":"Safety Monitors Mostly Catch What the Model Already Refuses","created":"2026-09-05","cited_by_count":0,"area":"safety"},{"id":"2609.10594","title":"An Empirical Measurement of Jailbreaking Evaluators","created":"2026-09-07","cited_by_count":0,"area":"safety"},{"id":"2609.36657","title":"Constitutional adapters: Inference-time interventions for misalignment and misuse","created":"2026-09-29","cited_by_count":0,"area":"safety"},{"id":"2610.05282","title":"Red-TTT: Test-Time Training for Automated Jailbreaking Large Language Models","created":"2026-10-04","cited_by_count":0,"area":"safety"}],"next_cursor":null}