{"total":64,"rows":[{"id":"2503.03704","title":"Memory Injection Attacks on LLM Agents via Query-Only Interaction","created":"2025-03-05","cited_by_count":144,"area":"agents"},{"id":"2502.17420","title":"The Geometry of Refusal in Large Language Models: Concept Cones and Representational Independence","created":"2025-02-24","cited_by_count":115,"area":"alignment"},{"id":"2507.11878","title":"LLMs Encode Harmfulness and Refusal Separately","created":"2025-07-16","cited_by_count":88,"area":"safety"},{"id":"2506.22355","title":"Embodied AI Agents: Modeling the World","created":"2025-06-27","cited_by_count":61,"area":"wm"},{"id":"2503.14477","title":"Calibrating Verbal Uncertainty as a Linear Feature to Reduce Hallucinations","created":"2025-03-18","cited_by_count":48,"area":null},{"id":"2502.07218","title":"LLM Unlearning via Neural Activation Redirection","created":"2025-02-11","cited_by_count":41,"area":"safety"},{"id":"2504.01550","title":"Representation Bending for Large Language Model Safety","created":"2025-04-02","cited_by_count":30,"area":"safety"},{"id":"2503.23084","title":"The Reasoning-Memorization Interplay in Language Models Is Mediated by a Single Direction","created":"2025-03-29","cited_by_count":21,"area":"lr"},{"id":"2601.03555","title":"SCRIBE: Structured Mid-Level Supervision for Tool-Using Language Models","created":"2026-01-07","cited_by_count":21,"area":"alignment"},{"id":"2506.00085","title":"COSMIC: Generalized Refusal Direction Identification in LLM Activations","created":"2025-05-30","cited_by_count":19,"area":null},{"id":"2503.02574","title":"LLM-Safety Evaluations Lack Robustness","created":"2025-03-04","cited_by_count":18,"area":"alignment"},{"id":"2602.11865","title":"Intelligent AI Delegation","created":"2026-02-12","cited_by_count":18,"area":null},{"id":"2603.19469","title":"A Framework for Formalizing LLM Agent Security","created":"2026-03-19","cited_by_count":16,"area":"agents"},{"id":"2509.23292","title":"Learning How to Use Tools, Not Just When: Pattern-Aware Tool-Integrated Reasoning","created":"2025-09-27","cited_by_count":15,"area":"lr"},{"id":"2502.04204","title":"Short-length Adversarial Training Helps LLMs Defend Long-length Jailbreak Attacks: Theoretical and Empirical Evidence","created":"2025-02-06","cited_by_count":14,"area":"safety"},{"id":"2601.22149","title":"DynaWeb: Model-Based Reinforcement Learning of Web Agents","created":"2026-01-29","cited_by_count":14,"area":"wm"},{"id":"2509.13281","title":"RepIt: Steering Language Models with Concept-Specific Refusal Vectors","created":"2025-09-16","cited_by_count":12,"area":null},{"id":"2510.01270","title":"Think Twice, Generate Once: Safeguarding by Progressive Self-Reflection","created":"2025-09-29","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":"2504.03770","title":"JailDAM: Jailbreak Detection with Adaptive Memory for Vision-Language Model","created":"2025-04-03","cited_by_count":10,"area":"mllm"},{"id":"2505.20309","title":"Guiding Giants: Lightweight Controllers for Weighted Activation Steering in LLMs","created":"2025-05-22","cited_by_count":8,"area":"mechinterp"},{"id":"2506.11938","title":"Improving Large Language Model Safety with Contrastive Representation Learning","created":"2025-06-13","cited_by_count":8,"area":"mechinterp"},{"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":"2511.10714","title":"BadThink: Triggered Overthinking Attacks on Chain-of-Thought Reasoning in Large Language Models","created":"2025-11-13","cited_by_count":8,"area":"lr"},{"id":"2602.00333","title":"Efficient and accurate steering of Large Language Models through attention-guided feature learning","created":"2026-01-30","cited_by_count":8,"area":null},{"id":"2501.10639","title":"Latent-space adversarial training with post-aware calibration for defending large language models against jailbreak attacks","created":"2025-01-18","cited_by_count":7,"area":"alignment"},{"id":"2604.08524","title":"What Drives Representation Steering? A Mechanistic Case Study on Steering Refusal","created":"2026-04-09","cited_by_count":7,"area":"alignment"},{"id":"2410.13691","title":"Jailbreaking LLM-Controlled Robots","created":"2024-10-17","cited_by_count":6,"area":"safety"},{"id":"2505.03096","title":"Assessing and Enhancing the Robustness of LLM-based Multi-Agent Systems Through Chaos Engineering","created":"2025-05-06","cited_by_count":6,"area":"agents"},{"id":"2507.17075","title":"LoRA is All You Need for Safety Alignment of Reasoning LLMs","created":"2025-07-22","cited_by_count":6,"area":"lr"},{"id":"2510.02324","title":"Hallucination reduction with CASAL: Contrastive Activation Steering For Amortized Learning","created":"2025-09-25","cited_by_count":6,"area":"moe"},{"id":"2412.08642","title":"Generative Semantic Communication: Architectures, Technologies, and Applications","created":"2024-12-11","cited_by_count":5,"area":null},{"id":"2502.04643","title":"Confidence Elicitation: A New Attack Vector for Large Language Models","created":"2025-02-07","cited_by_count":5,"area":null},{"id":"2505.09602","title":"Adversarial Suffix Filtering: a Defense Pipeline for LLMs","created":"2025-05-14","cited_by_count":5,"area":"safety"},{"id":"2508.20766","title":"Turning the Spell Around: Lightweight Alignment Amplification via Rank-One Safety Injection","created":"2025-08-28","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":"2504.03735","title":"Misaligned Roles, Misplaced Images: Structural Input Perturbations Expose Multimodal Alignment Blind Spots","created":"2025-04-01","cited_by_count":4,"area":"mllm"},{"id":"2506.07356","title":"Safety-Aligned Weights Are Not Enough: Refusal-Teacher-Guided Finetuning Enhances Safety and Downstream Performance under Harmful Finetuning Attacks","created":"2025-06-09","cited_by_count":2,"area":"safety"},{"id":"2602.06911","title":"TamperBench: Systematically Stress-Testing LLM Safety Under Fine-Tuning and Tampering","created":"2026-02-06","cited_by_count":2,"area":"safety"},{"id":"2602.16977","title":"Fail-Closed Alignment for Large Language Models","created":"2026-02-19","cited_by_count":2,"area":"alignment"},{"id":"2604.27019","title":"Dynamic Adversarial Fine-Tuning Reorganizes Refusal Geometry","created":"2026-04-29","cited_by_count":2,"area":"alignment"},{"id":"2605.06901","title":"Reflections and New Directions for Human-Centered Large Language Models","created":"2026-05-07","cited_by_count":2,"area":null},{"id":"2509.25843","title":"ASGuard: Activation-Scaling Guard to Mitigate Targeted Jailbreaking Attack","created":"2025-09-30","cited_by_count":1,"area":"safety"},{"id":"2511.07107","title":"MENTOR: A Metacognition-Driven Self-Evolution Framework for Uncovering and Mitigating Implicit Domain Risks in LLMs","created":"2025-11-10","cited_by_count":1,"area":"safety"},{"id":"2512.04144","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","created":"2025-12-03","cited_by_count":1,"area":"safety"},{"id":"2603.04355","title":"Efficient Refusal Ablation in LLM through Optimal Transport","created":"2026-03-04","cited_by_count":1,"area":"alignment"},{"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":"2603.14825","title":"Two Birds, One Projection: Harmonizing Safety and Utility in LVLMs via Inference-time Feature Projection","created":"2026-03-16","cited_by_count":1,"area":"mllm"},{"id":"2604.12817","title":"Understanding and Improving Continuous Adversarial Training for LLMs via In-context Learning Theory","created":"2026-04-14","cited_by_count":1,"area":"safety"},{"id":"2605.01913","title":"RefusalGuard: Geometry-Preserving Fine-Tuning for Safety in LLMs","created":"2026-05-03","cited_by_count":1,"area":"alignment"},{"id":"2605.08496","title":"Latent Personality Alignment: Improving Harmlessness Without Mentioning Harms","created":"2026-05-08","cited_by_count":1,"area":"safety"},{"id":"2605.16600","title":"Where Pretraining writes and Alignment reads: the asymmetry of Transformer weight space","created":"2026-05-15","cited_by_count":1,"area":"alignment"},{"id":"2608.24988","title":"Does Fine-Tuning Undo Activation Steering? Behavioural Recovery Without Weight-Edit Reversal","created":"2026-08-25","cited_by_count":1,"area":"mechinterp"},{"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":"2602.15238","title":"Closing the Distribution Gap in Adversarial Training for LLMs","created":"2026-02-16","cited_by_count":0,"area":null},{"id":"2604.00324","title":"The Persistent Vulnerability of Aligned AI Systems","created":"2026-03-31","cited_by_count":0,"area":null},{"id":"2605.24550","title":"Jailbreak to Protect: Buffering and Reinforcing via Temporary Jailbreaking for Safe Fine-Tuning in Large Language Models","created":"2026-05-23","cited_by_count":0,"area":"safety"},{"id":"2605.29816","title":"Harnessing non-adversarial robustness in large language models","created":"2026-05-28","cited_by_count":0,"area":null},{"id":"2606.07970","title":"Defending Against Malicious Finetuning by Scaling Train-time Adversarial Attacks","created":"2026-06-06","cited_by_count":0,"area":"alignment"},{"id":"2606.16349","title":"From Refusal Geometry to Safety Geometry: Harmfulness--Refusal Coupling under Dynamic Adversarial Fine-Tuning","created":"2026-06-15","cited_by_count":0,"area":"alignment"},{"id":"2606.25836","title":"AI Snitches Get Glitches: Towards Evading Agentic Surveillance","created":"2026-06-24","cited_by_count":0,"area":"agents"},{"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":"2608.25390","title":"Refusal geometry reflects refusal training: diverse refusal prefixes can raise stable rank and weaken refusal vector ablation attacks","created":"2026-08-26","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"}],"next_cursor":null}