{"total":390,"rows":[{"id":"2411.04468","title":"Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks","created":"2024-11-07","cited_by_count":232,"area":"agents"},{"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":"2504.15585","title":"A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deployment","created":"2025-04-22","cited_by_count":116,"area":"lr"},{"id":"2503.04957","title":"SafeArena: Evaluating the Safety of Autonomous Web Agents","created":"2025-03-06","cited_by_count":102,"area":"agents"},{"id":"2503.09648","title":"A Survey on Trustworthy LLM Agents: Threats and Countermeasures","created":"2025-03-12","cited_by_count":96,"area":"agents"},{"id":"2411.11296","title":"Steering Language Model Refusal with Sparse Autoencoders","created":"2024-11-18","cited_by_count":95,"area":"mechinterp"},{"id":"2410.10700","title":"LLMs know their vulnerabilities: Uncover Safety Gaps through Natural Distribution Shifts","created":"2024-10-14","cited_by_count":86,"area":null},{"id":"2410.02832","title":"FlipAttack: Jailbreak LLMs via Flipping","created":"2024-10-02","cited_by_count":80,"area":"safety"},{"id":"2407.11969","title":"Does Refusal Training in LLMs Generalize to the Past Tense?","created":"2024-07-16","cited_by_count":79,"area":"safety"},{"id":"2410.13886","title":"Refusal-Trained LLMs Are Easily Jailbroken As Browser Agents","created":"2024-10-11","cited_by_count":65,"area":"agents"},{"id":"2504.17704","title":"Safety in Large Reasoning Models: A Survey","created":"2025-04-24","cited_by_count":64,"area":"lr"},{"id":"2408.04686","title":"Multi-Turn Context Jailbreak Attack on Large Language Models From First Principles","created":"2024-08-08","cited_by_count":56,"area":"safety"},{"id":"2502.19820","title":"Foot-In-The-Door: A Multi-turn Jailbreak for LLMs","created":"2025-02-27","cited_by_count":54,"area":"safety"},{"id":"2412.07097","title":"On Evaluating the Durability of Safeguards for Open-Weight LLMs","created":"2024-12-10","cited_by_count":52,"area":null},{"id":"2406.09324","title":"Bag of Tricks: Benchmarking of Jailbreak Attacks on LLMs","created":"2024-06-13","cited_by_count":51,"area":"safety"},{"id":"2410.12855","title":"JAILJUDGE: A Comprehensive Jailbreak Judge Benchmark with Multi-Agent Enhanced Explanation Evaluation Framework","created":"2024-10-11","cited_by_count":47,"area":"agents"},{"id":"2410.01606","title":"Automated Red Teaming with GOAT: the Generative Offensive Agent Tester","created":"2024-10-02","cited_by_count":45,"area":"agents"},{"id":"2503.03710","title":"Improving LLM Safety Alignment with Dual-Objective Optimization","created":"2025-03-05","cited_by_count":43,"area":"alignment"},{"id":"2503.05021","title":"Safety is Not Only About Refusal: Reasoning-Enhanced Fine-tuning for Interpretable LLM Safety","created":"2025-03-06","cited_by_count":39,"area":"lr"},{"id":"2602.16901","title":"AgentLAB: Benchmarking LLM Agents against Long-Horizon Attacks","created":"2026-02-18","cited_by_count":39,"area":"agents"},{"id":"2411.14110","title":"Feedback-Guided Extraction of Knowledge Base from Retrieval-Augmented LLM Applications","created":"2024-11-21","cited_by_count":37,"area":"rag"},{"id":"2503.07885","title":"Safety Guardrails for LLM-Enabled Robots","created":"2025-03-10","cited_by_count":35,"area":"safety"},{"id":"2410.15236","title":"Jailbreaking and Mitigation of Vulnerabilities in Large Language Models","created":"2024-10-20","cited_by_count":34,"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":"2411.03814","title":"MRJ-Agent: An Effective Jailbreak Agent for Multi-Round Dialogue","created":"2024-11-06","cited_by_count":33,"area":"agents"},{"id":"2509.13782","title":"Who is Introducing the Failure? Automatically Attributing Failures of Multi-Agent Systems via Spectrum Analysis","created":"2025-09-17","cited_by_count":33,"area":"agents"},{"id":"2407.13833","title":"Phi-3 Safety Post-Training: Aligning Language Models with a \"Break-Fix\" Cycle","created":"2024-07-18","cited_by_count":32,"area":"alignment"},{"id":"2506.03350","title":"Adversarial Attacks on Robotic Vision Language Action Models","created":"2025-06-03","cited_by_count":32,"area":"vla"},{"id":"2406.09321","title":"JailbreakEval: An Integrated Toolkit for Evaluating Jailbreak Attempts Against Large Language Models","created":"2024-06-13","cited_by_count":31,"area":"safety"},{"id":"2410.02828","title":"PyRIT: A Framework for Security Risk Identification and Red Teaming in Generative AI System","created":"2024-10-01","cited_by_count":31,"area":"safety"},{"id":"2503.01829","title":"Persuade Me if You Can: A Framework for Evaluating Persuasion Effectiveness and Susceptibility Among Large Language Models","created":"2025-03-03","cited_by_count":27,"area":null},{"id":"2604.11790","title":"ClawGuard: A Runtime Security Framework for Tool-Augmented LLM Agents Against Indirect Prompt Injection","created":"2026-04-13","cited_by_count":27,"area":"agents"},{"id":"2604.16548","title":"A Survey on Long-Term Memory Security in LLM Agents: Attacks, Defenses, and Governance Across the Memory Lifecycle","created":"2026-04-17","cited_by_count":27,"area":"agents"},{"id":"2406.06622","title":"Adversarial Tuning: Defending Against Jailbreak Attacks for LLMs","created":"2024-06-07","cited_by_count":26,"area":"safety"},{"id":"2506.11094","title":"The Scales of Justitia: A Comprehensive Survey on Safety Evaluation of LLMs","created":"2025-06-06","cited_by_count":26,"area":null},{"id":"2410.15362","title":"Faster-GCG: Efficient Discrete Optimization Jailbreak Attacks against Aligned Large Language Models","created":"2024-10-20","cited_by_count":25,"area":"alignment"},{"id":"2507.02844","title":"Visual Contextual Attack: Jailbreaking MLLMs with Image-Driven Context Injection","created":"2025-07-03","cited_by_count":25,"area":"mllm"},{"id":"2506.14922","title":"FORTRESS: Frontier Risk Evaluation for National Security and Public Safety","created":"2025-06-17","cited_by_count":24,"area":"safety"},{"id":"2501.07238","title":"Lessons From Red Teaming 100 Generative AI Products","created":"2025-01-13","cited_by_count":23,"area":"safety"},{"id":"2502.11090","title":"SafeDialBench: A Fine-Grained Safety Evaluation Benchmark for Large Language Models in Multi-Turn Dialogues with Diverse Jailbreak Attacks","created":"2025-02-16","cited_by_count":23,"area":"safety"},{"id":"2507.00971","title":"Reasoning as an Adaptive Defense for Safety","created":"2025-07-01","cited_by_count":23,"area":"lr"},{"id":"2510.21236","title":"AgentBound: Securing Execution Boundaries of AI Agents","created":"2025-10-24","cited_by_count":23,"area":"agents"},{"id":"2410.16950","title":"Breaking ReAct Agents: Foot-in-the-Door Attack Will Get You In","created":"2024-10-22","cited_by_count":22,"area":"agents"},{"id":"2506.10597","title":"SoK: Evaluating Jailbreak Guardrails for Large Language Models","created":"2025-06-12","cited_by_count":21,"area":"alignment"},{"id":"2602.09222","title":"MUZZLE: Adaptive Agentic Red-Teaming of Web Agents Against Indirect Prompt Injection Attacks","created":"2026-02-09","cited_by_count":21,"area":"agents"},{"id":"2406.07954","title":"Dataset and Lessons Learned from the 2024 SaTML LLM Capture-the-Flag Competition","created":"2024-06-12","cited_by_count":20,"area":null},{"id":"2502.09990","title":"X-Boundary: Establishing Exact Safety Boundary to Shield LLMs from Multi-Turn Jailbreaks without Compromising Usability","created":"2025-02-14","cited_by_count":20,"area":"alignment"},{"id":"2503.00187","title":"Steering Dialogue Dynamics for Robustness against Multi-turn Jailbreaking Attacks","created":"2025-02-28","cited_by_count":20,"area":"safety"},{"id":"2505.16186","title":"SafeKey: Amplifying Aha-Moment Insights for Safety Reasoning","created":"2025-05-22","cited_by_count":20,"area":"lr"},{"id":"2604.04759","title":"Your Agent, Their Asset: A Real-World Safety Analysis of OpenClaw","created":"2026-04-06","cited_by_count":20,"area":"agents"},{"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":"2410.01294","title":"Endless Jailbreaks with Bijection Learning","created":"2024-10-02","cited_by_count":19,"area":"safety"},{"id":"2410.11459","title":"Jigsaw Puzzles: Splitting Harmful Questions to Jailbreak Large Language Models","created":"2024-10-15","cited_by_count":19,"area":"lr"},{"id":"2411.06426","title":"SequentialBreak: Large Language Models Can be Fooled by Embedding Jailbreak Prompts into Sequential Prompt Chains","created":"2024-11-10","cited_by_count":19,"area":"safety"},{"id":"2407.02551","title":"Breach By A Thousand Leaks: Unsafe Information Leakage in `Safe' AI Responses","created":"2024-07-02","cited_by_count":18,"area":"safety"},{"id":"2410.00296","title":"VLMGuard: Bootstrapping Malicious Prompt Detectors from Unlabeled Vision-Language Prompts in the Wild","created":"2024-10-01","cited_by_count":18,"area":"mllm"},{"id":"2508.07173","title":"Omni-SafetyBench: A Benchmark for Safety Evaluation of Audio-Visual Large Language Models","created":"2025-08-10","cited_by_count":18,"area":"safety"},{"id":"2503.10619","title":"Tempest: Autonomous Multi-Turn Jailbreaking of Large Language Models with Tree Search","created":"2025-03-13","cited_by_count":17,"area":"safety"},{"id":"2508.07646","title":"Multi-Turn Jailbreaks Are Simpler Than They Seem","created":"2025-08-11","cited_by_count":17,"area":"safety"},{"id":"2409.00137","title":"Emerging Vulnerabilities in Frontier Models: Multi-Turn Jailbreak Attacks","created":"2024-08-29","cited_by_count":16,"area":"safety"},{"id":"2511.05359","title":"ConVerse: Benchmarking Contextual Safety in Agent-to-Agent Conversations","created":"2025-11-07","cited_by_count":16,"area":"agents"},{"id":"2603.22928","title":"SoK: The Attack Surface of Agentic AI - Tools and Autonomy","created":"2026-03-24","cited_by_count":16,"area":"rag"},{"id":"2509.24496","title":"LLM DNA: Tracing Model Evolution via Functional Representations","created":"2025-09-29","cited_by_count":15,"area":null},{"id":"2406.03805","title":"AutoJailbreak: Exploring Jailbreak Attacks and Defenses through a Dependency Lens","created":"2024-06-06","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":"2512.07761","title":"TROJail: Trajectory-Level Optimization for Multi-Turn Large Language Model Jailbreaks with Process Rewards","created":"2025-12-08","cited_by_count":14,"area":"safety"},{"id":"2406.11978","title":"Dialogue Action Tokens: Steering Language Models in Goal-Directed Dialogue with a Multi-Turn Planner","created":"2024-06-17","cited_by_count":13,"area":"agents"},{"id":"2501.14250","title":"Siren: A Learning-Based Multi-Turn Attack Framework for Simulating Real-World Human Jailbreak Behaviors","created":"2025-01-24","cited_by_count":13,"area":"safety"},{"id":"2504.03174","title":"Multi-lingual Multi-turn Automated Red Teaming for LLMs","created":"2025-04-04","cited_by_count":13,"area":"safety"},{"id":"2508.04039","title":"Large Reasoning Models Are Autonomous Jailbreak Agents","created":"2025-08-04","cited_by_count":13,"area":"lr"},{"id":"2505.15389","title":"Are Vision-Language Models Safe in the Wild? A Meme-Based Benchmark Study","created":"2025-05-21","cited_by_count":12,"area":"mllm"},{"id":"2505.21605","title":"SoSBench: Benchmarking Safety Alignment on Six Scientific Domains","created":"2025-05-27","cited_by_count":12,"area":"alignment"},{"id":"2602.06854","title":"SEMA: Simple yet Effective Learning for Multi-Turn Jailbreak Attacks","created":"2026-02-06","cited_by_count":12,"area":"safety"},{"id":"2503.01742","title":"Building Safe GenAI Applications: An End-to-End Overview of Red Teaming for Large Language Models","created":"2025-03-03","cited_by_count":11,"area":"safety"},{"id":"2509.14285","title":"A Multi-Agent LLM Defense Pipeline Against Prompt Injection Attacks","created":"2025-09-16","cited_by_count":11,"area":"agents"},{"id":"2512.12069","title":"Rethinking Jailbreak Detection of Large Vision Language Models with Representational Contrastive Scoring","created":"2025-12-12","cited_by_count":11,"area":"lr"},{"id":"2503.06269","title":"Using Mechanistic Interpretability to Craft Adversarial Attacks against Large Language Models","created":"2025-03-08","cited_by_count":10,"area":"mechinterp"},{"id":"2503.19134","title":"MIRAGE: Multimodal Immersive Reasoning and Guided Exploration for Red-Team Jailbreak Attacks","created":"2025-03-24","cited_by_count":10,"area":"lr"},{"id":"2505.01050","title":"Transferable Adversarial Attacks on Black-Box Vision-Language Models","created":"2025-05-02","cited_by_count":10,"area":"mllm"},{"id":"2505.18471","title":"Invisible Tokens, Visible Bills: The Urgent Need to Audit Hidden Operations in Opaque LLM Services","created":"2025-05-24","cited_by_count":10,"area":null},{"id":"2507.02956","title":"A Representation Engineering Perspective on the Effectiveness of Multi-Turn Jailbreaks","created":"2025-06-29","cited_by_count":10,"area":"safety"},{"id":"2507.05248","title":"Response Attack: Exploiting Contextual Priming to Jailbreak Large Language Models","created":"2025-07-07","cited_by_count":10,"area":"safety"},{"id":"2510.00565","title":"Toward Safer Diffusion Language Models: Discovery and Mitigation of Priming Vulnerability","created":"2025-10-01","cited_by_count":10,"area":"diffusion"},{"id":"2510.07176","title":"Exposing LLM User Privacy via Traffic Fingerprint Analysis: A Study of Privacy Risks in LLM Agent Interactions","created":"2025-10-08","cited_by_count":10,"area":"agents"},{"id":"2502.09638","title":"Jailbreaking to Jailbreak","created":"2025-02-09","cited_by_count":9,"area":"safety"},{"id":"2505.14289","title":"EVA: Evolving Semantic Adversaries for Red-Teaming GUI Agents Against Environmental Injection Attacks","created":"2025-05-20","cited_by_count":9,"area":"agents"},{"id":"2505.17650","title":"Does Chain-of-Thought Reasoning Really Reduce Harmfulness from Jailbreaking?","created":"2025-05-23","cited_by_count":9,"area":"safety"},{"id":"2506.07524","title":"TAI3: Testing Agent Integrity in Interpreting User Intent","created":"2025-06-09","cited_by_count":9,"area":"agents"},{"id":"2507.22063","title":"RedCoder: Automated Multi-Turn Red Teaming for Code LLMs","created":"2025-06-25","cited_by_count":9,"area":"safety"},{"id":"2601.09625","title":"The Promptware Kill Chain: How Prompt Injections Gradually Evolved Into a Multistep Malware Delivery Mechanism","created":"2026-01-14","cited_by_count":9,"area":"safety"},{"id":"2502.08142","title":"Bridging the Safety Gap: A Guardrail Pipeline for Trustworthy LLM Inferences","created":"2025-02-12","cited_by_count":8,"area":"safety"},{"id":"2504.01278","title":"Strategize Globally, Adapt Locally: A Multi-Turn Red Teaming Agent with Dual-Level Learning","created":"2025-04-02","cited_by_count":8,"area":"agents"},{"id":"2506.06414","title":"Benchmarking Misuse Mitigation Against Covert Adversaries","created":"2025-06-06","cited_by_count":8,"area":null},{"id":"2507.01020","title":"AutoAdv: Automated Adversarial Prompting for Multi-Turn Jailbreaking of Large Language Models","created":"2025-04-18","cited_by_count":8,"area":"safety"},{"id":"2509.25624","title":"STAC: When Innocent Tools Form Dangerous Chains for LLM Agents","created":"2025-09-30","cited_by_count":8,"area":"agents"},{"id":"2510.02286","title":"Tree-based Dialogue Reinforced Policy Optimization for Red-Teaming Attacks","created":"2025-10-02","cited_by_count":8,"area":"safety"},{"id":"2510.22620","title":"Breaking Agent Backbones: Evaluating the Security of Backbone LLMs in AI Agents","created":"2025-10-26","cited_by_count":8,"area":"agents"},{"id":"2511.12710","title":"Evolve the Method, Not the Prompts: Evolutionary Synthesis of Jailbreak Attacks on LLMs","created":"2025-11-16","cited_by_count":8,"area":"safety"},{"id":"2605.01687","title":"MultiBreak: A Scalable and Diverse Multi-turn Jailbreak Benchmark for Evaluating LLM Safety","created":"2026-05-03","cited_by_count":8,"area":"safety"},{"id":"2407.02987","title":"LoRA-Guard: Parameter-Efficient Guardrail Adaptation for Content Moderation of Large Language Models","created":"2024-07-03","cited_by_count":7,"area":"alignment"}],"next_cursor":"100"}