{"total":79,"rows":[{"id":"2503.18666","title":"AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents","created":"2025-03-24","cited_by_count":223,"area":"agents"},{"id":"2408.17003","title":"Safety Layers in Aligned Large Language Models: The Key to LLM Security","created":"2024-08-30","cited_by_count":134,"area":"alignment"},{"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":"2410.13708","title":"On the Role of Attention Heads in Large Language Model Safety","created":"2024-10-17","cited_by_count":72,"area":"safety"},{"id":"2406.05498","title":"SelfDefend: LLMs Can Defend Themselves against Jailbreaking in a Practical Manner","created":"2024-06-08","cited_by_count":63,"area":"safety"},{"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":"2502.07557","title":"JBShield: Defending Large Language Models from Jailbreak Attacks through Activated Concept Analysis and Manipulation","created":"2025-02-11","cited_by_count":54,"area":"safety"},{"id":"2406.09324","title":"Bag of Tricks: Benchmarking of Jailbreak Attacks on LLMs","created":"2024-06-13","cited_by_count":51,"area":"safety"},{"id":"2407.14937","title":"Operationalizing a Threat Model for Red-Teaming Large Language Models (LLMs)","created":"2024-07-20","cited_by_count":46,"area":"safety"},{"id":"2502.14744","title":"HiddenDetect: Detecting Jailbreak Attacks against Large Vision-Language Models via Monitoring Hidden States","created":"2025-02-20","cited_by_count":33,"area":"mllm"},{"id":"2412.17034","title":"Shaping the Safety Boundaries: Understanding and Defending Against Jailbreaks in Large Language Models","created":"2024-12-22","cited_by_count":23,"area":"safety"},{"id":"2509.11864","title":"NeuroStrike: Neuron-Level Attacks on Aligned LLMs","created":"2025-09-15","cited_by_count":22,"area":"alignment"},{"id":"2506.10597","title":"SoK: Evaluating Jailbreak Guardrails for Large Language Models","created":"2025-06-12","cited_by_count":21,"area":"alignment"},{"id":"2410.01174","title":"Towards Inference-time Category-wise Safety Steering for Large Language Models","created":"2024-10-02","cited_by_count":18,"area":"alignment"},{"id":"2505.15404","title":"How Should We Enhance the Safety of Large Reasoning Models: An Empirical Study","created":"2025-05-21","cited_by_count":18,"area":"lr"},{"id":"2502.13095","title":"Understanding and Rectifying Safety Perception Distortion in VLMs","created":"2025-02-18","cited_by_count":17,"area":"mllm"},{"id":"2502.02958","title":"Position: Editing Large Language Models Poses Serious Safety Risks","created":"2025-02-05","cited_by_count":15,"area":null},{"id":"2502.01208","title":"On Almost Surely Safe Alignment of Large Language Models at Inference-Time","created":"2025-02-03","cited_by_count":14,"area":"alignment"},{"id":"2502.16776","title":"AISafetyLab: A Comprehensive Framework for AI Safety Evaluation and Improvement","created":"2025-02-24","cited_by_count":12,"area":null},{"id":"2512.06914","title":"SoK: Trust-Authorization Mismatch in LLM Agent Interactions","created":"2025-12-07","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":"2604.18519","title":"LLM Safety From Within: Detecting Harmful Content with Internal Representations","created":"2026-04-20","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":"2506.04250","title":"SafeSteer: Interpretable Safety Steering with Refusal-Evasion in LLMs","created":"2025-06-01","cited_by_count":10,"area":"mechinterp"},{"id":"2506.05451","title":"Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety","created":"2025-06-05","cited_by_count":10,"area":"safety"},{"id":"2410.09097","title":"Recent advancements in LLM Red-Teaming: Techniques, Defenses, and Ethical Considerations","created":"2024-10-09","cited_by_count":8,"area":"safety"},{"id":"2410.11182","title":"A Middle Path for On-Premises LLM Deployment: Preserving Privacy Without Sacrificing Model Confidentiality","created":"2024-10-15","cited_by_count":8,"area":null},{"id":"2506.11938","title":"Improving Large Language Model Safety with Contrastive Representation Learning","created":"2025-06-13","cited_by_count":8,"area":"mechinterp"},{"id":"2601.23255","title":"Now You Hear Me: Audio Narrative Attacks Against Large Audio-Language Models","created":"2026-01-30","cited_by_count":8,"area":null},{"id":"2501.02629","title":"Layer-Level Self-Exposure and Patch: Affirmative Token Mitigation for Jailbreak Attack Defense","created":"2025-01-05","cited_by_count":7,"area":"safety"},{"id":"2509.01631","title":"Unraveling LLM Jailbreaks Through Safety Knowledge Neurons","created":"2025-09-01","cited_by_count":7,"area":"safety"},{"id":"2408.08924","title":"Prefix Guidance: A Steering Wheel for Large Language Models to Defend Against Jailbreak Attacks","created":"2024-08-15","cited_by_count":6,"area":"safety"},{"id":"2502.20186","title":"Layer-Aware Task Arithmetic: Disentangling Task-Specific and Instruction-Following Knowledge","created":"2025-02-27","cited_by_count":6,"area":null},{"id":"2503.00037","title":"Zero-Shot Defense Against Toxic Images via Inherent Multimodal Alignment in LVLMs","created":"2025-02-25","cited_by_count":6,"area":"mllm"},{"id":"2505.14300","title":"Beyond Black-Box Obfuscation: Mechanistic Analysis and Defense of White-Box Monitors","created":"2025-05-20","cited_by_count":6,"area":null},{"id":"2505.14368","title":"From ASR to ASP: Evaluating Prompt Attack Vulnerabilities Against Open-Source LLMs","created":"2025-05-20","cited_by_count":6,"area":"safety"},{"id":"2506.00166","title":"Disentangled Safety Adapters Enable Efficient Guardrails and Flexible Inference-Time Alignment","created":"2025-05-30","cited_by_count":6,"area":"alignment"},{"id":"2505.24621","title":"Benchmarking Large Language Models for Cryptanalysis and Side-Channel Vulnerabilities","created":"2025-05-30","cited_by_count":5,"area":null},{"id":"2507.06043","title":"CAVGAN: Unifying Jailbreak and Defense of LLMs via Generative Adversarial Attacks on their Internal Representations","created":"2025-07-08","cited_by_count":5,"area":"safety"},{"id":"2510.10161","title":"Large Language Model Sourcing: A Survey","created":"2025-10-11","cited_by_count":5,"area":null},{"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":"2502.15594","title":"SafeInt: Shielding Large Language Models from Jailbreak Attacks via Safety-Aware Representation Intervention","created":"2025-02-21","cited_by_count":4,"area":"safety"},{"id":"2506.00382","title":"Spectral Insights into Data-Oblivious Critical Layers in Large Language Models","created":"2025-05-31","cited_by_count":4,"area":null},{"id":"2503.17932","title":"STShield: Single-Token Sentinel for Real-Time Jailbreak Detection in Large Language Models","created":"2025-03-23","cited_by_count":3,"area":"safety"},{"id":"2601.03600","title":"ALERT: Zero-shot LLM Jailbreak Detection via Internal Discrepancy Amplification","created":"2026-01-07","cited_by_count":3,"area":"safety"},{"id":"2601.05339","title":"Multi-turn Jailbreaking Attack in Multi-Modal Large Language Models","created":"2026-01-08","cited_by_count":3,"area":"safety"},{"id":"2604.09544","title":"Large Language Models Generate Harmful Responses Using a Distinct Mechanism, Shared Across Harm Types","created":"2026-04-10","cited_by_count":3,"area":"safety"},{"id":"2605.11002","title":"MT-JailBench: A Modular Benchmark for Understanding Multi-Turn Jailbreak Attacks","created":"2026-05-10","cited_by_count":3,"area":"safety"},{"id":"2508.01306","title":"PUZZLED: Jailbreaking LLMs through Word-Based Puzzles","created":"2025-08-02","cited_by_count":2,"area":"safety"},{"id":"2511.16229","title":"Q-MLLM: Vector Quantization for Robust Multimodal Large Language Model Security","created":"2025-11-20","cited_by_count":2,"area":"mllm"},{"id":"2601.22620","title":"Layer-wise Swapping for Generalizable Multilingual Safety","created":"2026-01-30","cited_by_count":2,"area":"safety"},{"id":"2604.27019","title":"Dynamic Adversarial Fine-Tuning Reorganizes Refusal Geometry","created":"2026-04-29","cited_by_count":2,"area":"alignment"},{"id":"2607.01250","title":"Structuring the Space of Sociotechnical Alignment","created":"2026-06-02","cited_by_count":2,"area":null},{"id":"2608.23959","title":"NeuronGuard: Robust LLM Safety Alignment via Ablation-Aware Safety Signal Redistribution","created":"2026-08-25","cited_by_count":2,"area":"alignment"},{"id":"2408.11308","title":"Defending against Jailbreak through Early Exit Generation of Large Language Models","created":"2024-08-21","cited_by_count":1,"area":"inference"},{"id":"2410.00454","title":"UniAdapt: A Universal Adapter for Knowledge Calibration","created":"2024-10-01","cited_by_count":1,"area":null},{"id":"2412.20476","title":"Cut the Deadwood Out: Backdoor Purification via Guided Module Substitution","created":"2024-12-29","cited_by_count":1,"area":null},{"id":"2603.05772","title":"Depth Charge: Jailbreak Large Language Models from Deep Safety Attention Heads","created":"2026-03-06","cited_by_count":1,"area":"safety"},{"id":"2605.07250","title":"Hard to Read, Easy to Jailbreak: How Visual Degradation Bypasses MLLM Safety Alignment","created":"2026-05-08","cited_by_count":1,"area":"alignment"},{"id":"2606.02640","title":"D-Judge: Disrupting Multi-Turn Jailbreaks using Semantics-Preserving Output Rewriting","created":"2026-05-31","cited_by_count":1,"area":"safety"},{"id":"2607.12406","title":"Isolation as a First-Class Principle for LLM-Agent System Safety: Concepts, Taxonomy, Challenges and Future Directions","created":"2026-07-14","cited_by_count":1,"area":"agents"},{"id":"2608.07430","title":"Diffusion LLMs as Targets and Adversaries: Mechanistic Safety Exploits","created":"2026-08-07","cited_by_count":1,"area":"safety"},{"id":"2505.07167","title":"Decoding One Safety Trigger Token for Balancing Safety and Usability in Large Language Models","created":"2025-05-12","cited_by_count":0,"area":"safety"},{"id":"2508.00602","title":"LeakSealer: A Semisupervised Defense for LLMs Against Prompt Injection and Leakage Attacks","created":"2025-08-01","cited_by_count":0,"area":"safety"},{"id":"2508.11290","title":"SafeConstellations: Mitigating Over-Refusals in LLMs Through Task-Aware Representation Steering","created":"2025-08-15","cited_by_count":0,"area":"safety"},{"id":"2512.04841","title":"SoK: a Comprehensive Causality Analysis Framework for Large Language Model Security","created":"2025-12-04","cited_by_count":0,"area":"safety"},{"id":"2601.03005","title":"JPU: Bridging Jailbreak Defense and Unlearning via On-Policy Path Rectification","created":"2026-01-06","cited_by_count":0,"area":"alignment"},{"id":"2602.06630","title":"TrapSuffix: Proactive Defense Against Adversarial Suffixes in Jailbreaking","created":"2026-02-06","cited_by_count":0,"area":"safety"},{"id":"2603.06727","title":"Safe Transformer: An Explicit Safety Bit For Interpretable And Controllable Alignment","created":"2026-03-06","cited_by_count":0,"area":"alignment"},{"id":"2603.15397","title":"SFCoT: Safer Chain-of-Thought via Active Safety Evaluation and Calibration","created":"2026-03-16","cited_by_count":0,"area":"lr"},{"id":"2604.07835","title":"Silencing the Guardrails: Inference-Time Jailbreaking via Dynamic Contextual Representation Ablation","created":"2026-04-09","cited_by_count":0,"area":"safety"},{"id":"2604.10031","title":"CoSToM:Causal-oriented Steering for Intrinsic Theory-of-Mind Alignment in Large Language Models","created":"2026-04-11","cited_by_count":0,"area":null},{"id":"2605.10582","title":"Guaranteed Jailbreaking Defense via Disrupt-and-Rectify Smoothing","created":"2026-05-11","cited_by_count":0,"area":"safety"},{"id":"2605.14750","title":"EVA: Editing for Versatile Alignment against Jailbreaks","created":"2026-05-14","cited_by_count":0,"area":"mllm"},{"id":"2605.19321","title":"Exploring and Developing a Pre-Model Safeguard with Draft Models","created":"2026-05-19","cited_by_count":0,"area":"safety"},{"id":"2608.11583","title":"Localizing Safety Alignment: MLP Layers and Mid-Network Blocks Encode Refusal Behavior in Large Language Models","created":"2026-08-12","cited_by_count":0,"area":"alignment"},{"id":"2608.14392","title":"Tripwire: Triggering Aligned Refusal via Statistically Certified Safety Neurons","created":"2026-08-14","cited_by_count":0,"area":"safety"},{"id":"2609.05850","title":"SAFEGuard: Detect Optimization-Based Jailbreak Attacks Through Harmful Semantic Analysis and Fluency Measurement","created":"2026-09-05","cited_by_count":0,"area":"alignment"},{"id":"2610.01058","title":"MOMAT: Mixture of Multiple Atlases for Low-Power Jailbreak Defense of Quantized LLMs","created":"2026-10-01","cited_by_count":0,"area":"safety"}],"next_cursor":null}