{"total":98,"rows":[{"id":"2607.15495","title":"Verbalizable Representations Form a Global Workspace in Language Models","created":"2026-07-16","cited_by_count":75,"area":null},{"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":"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":"2407.15017","title":"Knowledge Mechanisms in Large Language Models: A Survey and Perspective","created":"2024-07-22","cited_by_count":51,"area":null},{"id":"2506.07022","title":"AlphaSteer: Learning Refusal Steering with Principled Null-Space Constraint","created":"2025-06-08","cited_by_count":49,"area":"safety"},{"id":"2502.09674","title":"The Hidden Dimensions of LLM Alignment: A Multi-Dimensional Analysis of Orthogonal Safety Directions","created":"2025-02-13","cited_by_count":41,"area":"lr"},{"id":"2504.18041","title":"RAG LLMs are Not Safer: A Safety Analysis of Retrieval-Augmented Generation for Large Language Models","created":"2025-04-25","cited_by_count":24,"area":"rag"},{"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":"2407.07342","title":"Multilingual Blending: LLM Safety Alignment Evaluation with Language Mixture","created":"2024-07-10","cited_by_count":22,"area":"alignment"},{"id":"2411.11114","title":"JailbreakLens: Interpreting Jailbreak Mechanism in the Lens of Representation and Circuit","created":"2024-11-17","cited_by_count":22,"area":"safety"},{"id":"2509.03888","title":"False Sense of Security: Why Probing-based Malicious Input Detection Fails to Generalize","created":"2025-09-04","cited_by_count":19,"area":null},{"id":"2601.01685","title":"Lying with Truths: Open-Channel Multi-Agent Collusion for Belief Manipulation via Generative Montage","created":"2026-01-04","cited_by_count":19,"area":"lr"},{"id":"2506.00823","title":"Probing the Geometry of Truth: Consistency and Generalization of Truth Directions in LLMs Across Logical Transformations and Question Answering Tasks","created":"2025-06-01","cited_by_count":18,"area":null},{"id":"2507.16372","title":"Depth Gives a False Sense of Privacy: LLM Internal States Inversion","created":"2025-07-22","cited_by_count":17,"area":null},{"id":"2511.10400","title":"Rethinking the Reliability of Multi-agent System: A Perspective from Byzantine Fault Tolerance","created":"2025-11-13","cited_by_count":17,"area":"lr"},{"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":"2604.03147","title":"Valence-Arousal Subspace in LLMs: Circular Emotion Geometry and Multi-Behavioral Control","created":"2026-04-03","cited_by_count":16,"area":null},{"id":"2407.16637","title":"Course-Correction: Safety Alignment Using Synthetic Preferences","created":"2024-07-23","cited_by_count":13,"area":"alignment"},{"id":"2411.19530","title":"Quantized Delta Weight Is Safety Keeper","created":"2024-11-29","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":"2409.14195","title":"The Imperative of Conversation Analysis in the Era of LLMs: A Survey of Tasks, Techniques, and Trends","created":"2024-09-21","cited_by_count":11,"area":null},{"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":"2502.09755","title":"Jailbreak Attack Initializations as Extractors of Compliance Directions","created":"2025-02-13","cited_by_count":10,"area":"safety"},{"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":"2504.16489","title":"Amplified Vulnerabilities: Structured Jailbreak Attacks on LLM-based Multi-Agent Debate","created":"2025-04-23","cited_by_count":10,"area":"lr"},{"id":"2504.20965","title":"AegisLLM: Scaling Agentic Systems for Self-Reflective Defense in LLM Security","created":"2025-04-29","cited_by_count":10,"area":"agents"},{"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":"2507.20333","title":"The Blessing and Curse of Dimensionality in Safety Alignment","created":"2025-07-27","cited_by_count":10,"area":"alignment"},{"id":"2509.21761","title":"Backdoor Attribution: Elucidating and Controlling Backdoor in Language Models","created":"2025-09-26","cited_by_count":10,"area":"safety"},{"id":"2508.05775","title":"Guardians and Offenders: A Survey on Harmful Content Generation and Safety Mitigation of LLM","created":"2025-08-07","cited_by_count":9,"area":"alignment"},{"id":"2410.04524","title":"Toward Secure Tuning: Mitigating Security Risks from Instruction Fine-Tuning","created":"2024-10-06","cited_by_count":8,"area":null},{"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":"2505.23839","title":"GeneBreaker: Jailbreak Attacks against DNA Language Models with Pathogenicity Guidance","created":"2025-05-28","cited_by_count":8,"area":"safety"},{"id":"2509.15202","title":"Beyond Surface Alignment: Rebuilding LLMs Safety Mechanism via Probabilistically Ablating Refusal Direction","created":"2025-09-18","cited_by_count":8,"area":"alignment"},{"id":"2501.00055","title":"LLM-Virus: Evolutionary Jailbreak Attack on Large Language Models","created":"2024-12-28","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":"2601.17887","title":"When Personalization Legitimizes Risks: Uncovering Safety Vulnerabilities in Personalized Dialogue Agents","created":"2026-01-25","cited_by_count":7,"area":"agents"},{"id":"2502.12485","title":"Safe at the Margins: A General Approach to Safety Alignment in Low-Resource English Languages -- A Singlish Case Study","created":"2025-02-18","cited_by_count":6,"area":"alignment"},{"id":"2507.04531","title":"DP-Fusion: Token-Level Differentially Private Inference for Large Language Models","created":"2025-07-06","cited_by_count":6,"area":"safety"},{"id":"2512.10449","title":"When Reject Turns into Accept: Quantifying the Vulnerability of LLM-Based Scientific Reviewers to Indirect Prompt Injection","created":"2025-12-11","cited_by_count":6,"area":"safety"},{"id":"2505.22298","title":"Adaptive Detoxification: Safeguarding General Capabilities of LLMs through Toxicity-Aware Knowledge Editing","created":"2025-05-28","cited_by_count":5,"area":"safety"},{"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":"2511.00689","title":"Do Methods to Jailbreak and Defend LLMs Generalize Across Languages?","created":"2025-11-01","cited_by_count":5,"area":"alignment"},{"id":"2511.14423","title":"Unified Defense for Large Language Models against Jailbreak and Fine-Tuning Attacks in Education","created":"2025-11-18","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":"2505.05237","title":"Latte: Transfering LLMs` Latent-level Knowledge for Few-shot Tabular Learning","created":"2025-05-08","cited_by_count":4,"area":null},{"id":"2510.12264","title":"Reducing Belief Deviation in Reinforcement Learning for Active Reasoning","created":"2025-10-14","cited_by_count":4,"area":"lr"},{"id":"2603.23268","title":"SafeSeek: Universal Attribution of Safety Circuits in Language Models","created":"2026-03-24","cited_by_count":4,"area":"safety"},{"id":"2501.16534","title":"Targeting Alignment: Extracting Safety Classifiers of Aligned LLMs","created":"2025-01-27","cited_by_count":3,"area":"safety"},{"id":"2507.13474","title":"Paper Summary Attack: Jailbreaking LLMs through LLM Safety Papers","created":"2025-07-17","cited_by_count":3,"area":"safety"},{"id":"2511.20158","title":"Harmonious Parameter Adaptation in Continual Visual Instruction Tuning for Safety-Aligned MLLMs","created":"2025-11-25","cited_by_count":3,"area":"mllm"},{"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":"2602.04448","title":"RASA: Routing-Aware Safety Alignment for Mixture-of-Experts Models","created":"2026-02-04","cited_by_count":3,"area":"alignment"},{"id":"2605.05058","title":"SoK: Robustness in Large Language Models against Jailbreak Attacks","created":"2026-05-06","cited_by_count":3,"area":"safety"},{"id":"2605.12726","title":"Before the Last Token: Diagnosing Final-Token Safety Probe Failures","created":"2026-05-12","cited_by_count":3,"area":"safety"},{"id":"2408.07663","title":"Alignment-Enhanced Decoding:Defending via Token-Level Adaptive Refining of Probability Distributions","created":"2024-08-14","cited_by_count":2,"area":"alignment"},{"id":"2503.06989","title":"Probabilistic Modeling of Jailbreak on Multimodal LLMs: From Quantification to Application","created":"2025-03-10","cited_by_count":2,"area":"mllm"},{"id":"2510.17256","title":"Explainability of Large Language Models: Opportunities and Challenges toward Generating Trustworthy Explanations","created":"2025-10-20","cited_by_count":2,"area":"mechinterp"},{"id":"2602.01999","title":"From Latent Signals to Reflection Behavior: Tracing Meta-Cognitive Activation Trajectory in R1-Style LLMs","created":"2026-02-02","cited_by_count":2,"area":null},{"id":"2602.21346","title":"Alignment-Weighted DPO: A principled reasoning approach to improve safety alignment","created":"2026-02-24","cited_by_count":2,"area":"lr"},{"id":"2604.27019","title":"Dynamic Adversarial Fine-Tuning Reorganizes Refusal Geometry","created":"2026-04-29","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":"2412.20476","title":"Cut the Deadwood Out: Backdoor Purification via Guided Module Substitution","created":"2024-12-29","cited_by_count":1,"area":null},{"id":"2503.12931","title":"MirrorShield: Towards Universal Defense Against Jailbreaks via Entropy-Guided Mirror Crafting","created":"2025-03-17","cited_by_count":1,"area":"safety"},{"id":"2508.00555","title":"Activation-Guided Local Editing for Jailbreaking Attacks","created":"2025-08-01","cited_by_count":1,"area":"safety"},{"id":"2508.20038","title":"Forewarned is Forearmed: Pre-Synthesizing Jailbreak-like Instructions to Enhance LLM Safety Guardrail to Potential Attacks","created":"2025-08-27","cited_by_count":1,"area":"alignment"},{"id":"2509.03985","title":"NeuroBreak: Unveil Internal Jailbreak Mechanisms in Large Language Models","created":"2025-09-04","cited_by_count":1,"area":"alignment"},{"id":"2509.06807","title":"MoGU V2: Toward a Higher Pareto Frontier Between Model Usability and Security","created":"2025-09-08","cited_by_count":1,"area":null},{"id":"2511.19009","title":"Understanding and Mitigating Over-refusal for Large Language Models via Representation Intervention","created":"2025-11-24","cited_by_count":1,"area":"safety"},{"id":"2512.06655","title":"Graph-Regularized Sparse Autoencoders for LLM Safety Steering","created":"2025-12-07","cited_by_count":1,"area":"safety"},{"id":"2512.14751","title":"One Leak Away: How Pretrained Model Exposure Amplifies Jailbreak Risks in Finetuned LLMs","created":"2025-12-14","cited_by_count":1,"area":"safety"},{"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":"2603.28817","title":"GUARD-SLM: Token Activation-Based Defense Against Jailbreak Attacks for Small Language Models","created":"2026-03-28","cited_by_count":1,"area":"safety"},{"id":"2604.20930","title":"SafeRedirect: Defeating Internal Safety Collapse via Task-Completion Redirection in Frontier LLMs","created":"2026-04-22","cited_by_count":1,"area":"safety"},{"id":"2606.28153","title":"Robust Harmful Features Under Jailbreak Attacks: Mechanistic Evidence from Attention Head Specialization in Large Language Models","created":"2026-06-26","cited_by_count":1,"area":"alignment"},{"id":"2608.01117","title":"SoK: Intent-Oriented Systematization of Multi-Turn LLM Jailbreaks","created":"2026-08-02","cited_by_count":1,"area":"safety"},{"id":"2609.16204","title":"Decoy Direction Optimization: A Post-Hoc Defense Against LLM Abliteration","created":"2026-09-14","cited_by_count":1,"area":"safety"},{"id":"2502.20952","title":"Efficient Jailbreaking of Large Models by Freeze Training: Lower Layers Exhibit Greater Sensitivity to Harmful Content","created":"2025-02-28","cited_by_count":0,"area":"safety"},{"id":"2504.09191","title":"Feature-Aware Malicious Output Detection and Mitigation","created":"2025-04-12","cited_by_count":0,"area":null},{"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":"2506.23576","title":"Evaluating Multi-Agent Defences Against Jailbreaking Attacks on Large Language Models","created":"2025-06-30","cited_by_count":0,"area":"agents"},{"id":"2507.08014","title":"Mass-Scale Analysis of In-the-Wild Conversations Reveals Complexity Bounds on LLM Jailbreaking","created":"2025-07-06","cited_by_count":0,"area":"safety"},{"id":"2601.13288","title":"A BERTology View of LLM Orchestrations: Token- and Layer-Selective Probes for Efficient Single-Pass Classification","created":"2026-01-19","cited_by_count":0,"area":null},{"id":"2601.15801","title":"Attributing and Exploiting Safety Vectors through Global Optimization in Large Language Models","created":"2026-01-22","cited_by_count":0,"area":"safety"},{"id":"2603.15647","title":"Steering Frozen LLMs: Adaptive Social Alignment via Online Prompt Routing","created":"2026-03-04","cited_by_count":0,"area":"alignment"},{"id":"2604.06247","title":"SALLIE: Generation-Free Hidden-State Detection of Jailbreaks and Prompt Injections Across Text and Vision","created":"2026-04-06","cited_by_count":0,"area":"mllm"},{"id":"2604.11663","title":"Why Do Large Language Models Generate Harmful Content?","created":"2026-04-13","cited_by_count":0,"area":"safety"},{"id":"2605.02914","title":"When Safety Geometry Collapses: Fine-Tuning Vulnerabilities in Agentic Guard Models","created":"2026-04-08","cited_by_count":0,"area":"agents"},{"id":"2605.05630","title":"One Turn Too Late: Learning When to Intervene Against Multi-Turn Malicious Intent","created":"2026-05-07","cited_by_count":0,"area":null},{"id":"2605.08878","title":"Why Do Aligned LLMs Remain Jailbreakable: Refusal-Escape Directions, Operator-Level Sources, and Safety-Utility Trade-off","created":"2026-05-09","cited_by_count":0,"area":"alignment"},{"id":"2605.17971","title":"Babel: Jailbreaking Safety Attention via Obfuscation Distribution Optimized Sampling","created":"2026-05-18","cited_by_count":0,"area":"alignment"},{"id":"2605.30913","title":"Toxic HallucinAItions: Perturbing Prompts and Tracing LLM Circuits","created":"2026-05-29","cited_by_count":0,"area":null},{"id":"2606.19755","title":"SafeSpec: Fast and Safe LLM via Dynamic Reflective Sampling","created":"2026-06-18","cited_by_count":0,"area":"inference"},{"id":"2607.08423","title":"OmniFood-Bench: Evaluating VLMs for Nutrient Reasoning and Personalized Health Advice","created":"2026-07-09","cited_by_count":0,"area":"lr"},{"id":"2608.00828","title":"Isotropy Cliffs: The Geometric Signature of Decision-Making in Large Language Models","created":"2026-08-01","cited_by_count":0,"area":null},{"id":"2608.26506","title":"A Single Suffix to Break Them All: Basin-Aware Jailbreaks for Merged Model Families","created":"2026-08-27","cited_by_count":0,"area":"safety"},{"id":"2609.13737","title":"ForeSight: Enhancing Risk Monitoring via Early Safety Signal Distillation","created":"2026-09-12","cited_by_count":0,"area":"safety"},{"id":"2609.33039","title":"Agent Safety From Within: Detecting Harmful Trajectories from LLM Internal States","created":"2026-09-27","cited_by_count":0,"area":"agents"}],"next_cursor":null}