{"total":112,"rows":[{"id":"2403.13372","title":"LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models","created":"2024-03-20","cited_by_count":1464,"area":null},{"id":"2408.07666","title":"Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities","created":"2024-08-14","cited_by_count":254,"area":null},{"id":"2505.23836","title":"Large Language Models Often Know When They Are Being Evaluated","created":"2025-05-28","cited_by_count":128,"area":null},{"id":"2407.18418","title":"Know Your Limits: A Survey of Abstention in Large Language Models","created":"2024-07-25","cited_by_count":121,"area":null},{"id":"2409.01586","title":"Booster: Tackling Harmful Fine-tuning for Large Language Models via Attenuating Harmful Perturbation","created":"2024-09-03","cited_by_count":92,"area":null},{"id":"2409.18169","title":"Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey","created":"2024-09-26","cited_by_count":90,"area":"alignment"},{"id":"2507.11878","title":"LLMs Encode Harmfulness and Refusal Separately","created":"2025-07-16","cited_by_count":88,"area":"safety"},{"id":"2404.05399","title":"SafetyPrompts: a Systematic Review of Open Datasets for Evaluating and Improving Large Language Model Safety","created":"2024-04-08","cited_by_count":77,"area":"safety"},{"id":"2408.09600","title":"Antidote: Post-fine-tuning Safety Alignment for Large Language Models against Harmful Fine-tuning","created":"2024-08-18","cited_by_count":72,"area":"alignment"},{"id":"2405.09055","title":"A safety realignment framework via subspace-oriented model fusion for large language models","created":"2024-05-15","cited_by_count":41,"area":"safety"},{"id":"2402.16382","title":"Immunization against harmful fine-tuning attacks","created":"2024-02-26","cited_by_count":35,"area":null},{"id":"2406.11801","title":"Safety Arithmetic: A Framework for Test-time Safety Alignment of Language Models by Steering Parameters and Activations","created":"2024-06-17","cited_by_count":32,"area":"alignment"},{"id":"2505.17147","title":"MTSA: Multi-turn Safety Alignment for LLMs through Multi-round Red-teaming","created":"2025-05-22","cited_by_count":32,"area":"alignment"},{"id":"2405.16229","title":"No Two Devils Alike: Unveiling Distinct Mechanisms of Fine-tuning Attacks","created":"2024-05-25","cited_by_count":29,"area":"alignment"},{"id":"2505.23556","title":"Understanding Refusal in Language Models with Sparse Autoencoders","created":"2025-05-29","cited_by_count":29,"area":"alignment"},{"id":"2409.16783","title":"Holistic Automated Red Teaming for Large Language Models through Top-Down Test Case Generation and Multi-turn Interaction","created":"2024-09-25","cited_by_count":25,"area":"safety"},{"id":"2503.17239","title":"SafeMERGE: Preserving Safety Alignment in Fine-Tuned Large Language Models via Selective Layer-Wise Model Merging","created":"2025-03-21","cited_by_count":25,"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":"2411.18729","title":"Multi-Task Model Merging via Adaptive Weight Disentanglement","created":"2024-11-27","cited_by_count":22,"area":null},{"id":"2509.11864","title":"NeuroStrike: Neuron-Level Attacks on Aligned LLMs","created":"2025-09-15","cited_by_count":22,"area":"alignment"},{"id":"2412.10493","title":"AlignGuard: Scalable Safety Alignment for Text-to-Image Generation","created":"2024-12-13","cited_by_count":20,"area":"alignment"},{"id":"2412.11041","title":"Separate the Wheat from the Chaff: A Post-Hoc Approach to Safety Re-Alignment for Fine-Tuned Language Models","created":"2024-12-15","cited_by_count":20,"area":"alignment"},{"id":"2404.05188","title":"Have You Merged My Model? On The Robustness of Large Language Model IP Protection Methods Against Model Merging","created":"2024-04-08","cited_by_count":19,"area":"safety"},{"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":"2505.12038","title":"Safe Delta: Consistently Preserving Safety when Fine-Tuning LLMs on Diverse Datasets","created":"2025-05-17","cited_by_count":18,"area":null},{"id":"2405.13820","title":"Towards Comprehensive Post Safety Alignment of Large Language Models via Safety Patching","created":"2024-05-22","cited_by_count":17,"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":"2410.10801","title":"Mix Data or Merge Models? Optimizing for Diverse Multi-Task Learning","created":"2024-10-14","cited_by_count":17,"area":null},{"id":"2406.02925","title":"Task Arithmetic can Mitigate Synthetic-to-Real Gap in Automatic Speech Recognition","created":"2024-06-05","cited_by_count":14,"area":null},{"id":"2406.15796","title":"Unveiling Entity-Level Unlearning for Large Language Models: A Comprehensive Analysis","created":"2024-06-22","cited_by_count":14,"area":"safety"},{"id":"2410.10862","title":"Superficial Safety Alignment Hypothesis","created":"2024-10-07","cited_by_count":14,"area":"alignment"},{"id":"2410.05269","title":"Data Advisor: Dynamic Data Curation for Safety Alignment of Large Language Models","created":"2024-10-07","cited_by_count":13,"area":"alignment"},{"id":"2502.15924","title":"Improving Consistency in Large Language Models through Chain of Guidance","created":"2025-02-21","cited_by_count":13,"area":null},{"id":"2505.19056","title":"An Embarrassingly Simple Defense Against LLM Abliteration Attacks","created":"2025-05-25","cited_by_count":13,"area":"safety"},{"id":"2412.19512","title":"Safeguard Fine-Tuned LLMs Through Pre- and Post-Tuning Model Merging","created":"2024-12-27","cited_by_count":11,"area":null},{"id":"2504.15047","title":"RainbowPlus: Enhancing Adversarial Prompt Generation via Evolutionary Quality-Diversity Search","created":"2025-04-21","cited_by_count":11,"area":"safety"},{"id":"2512.21008","title":"GateBreaker: Gate-Guided Attacks on Mixture-of-Expert LLMs","created":"2025-12-24","cited_by_count":11,"area":"safety"},{"id":"2502.11244","title":"Soteria: Language-Specific Functional Parameter Steering for Multilingual Safety Alignment","created":"2025-02-16","cited_by_count":10,"area":"alignment"},{"id":"2506.04250","title":"SafeSteer: Interpretable Safety Steering with Refusal-Evasion in LLMs","created":"2025-06-01","cited_by_count":10,"area":"mechinterp"},{"id":"2602.12158","title":"SafeNeuron: Neuron-Level Safety Alignment for Large Language Models","created":"2026-02-12","cited_by_count":10,"area":"mllm"},{"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":"2505.16241","title":"Three Minds, One Legend: Jailbreak Large Reasoning Model with Adaptive Stacked Ciphers","created":"2025-05-22","cited_by_count":8,"area":"lr"},{"id":"2509.22745","title":"Defending MoE LLMs against Harmful Fine-Tuning via Safety Routing Alignment","created":"2025-09-26","cited_by_count":8,"area":"moe"},{"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":"2602.00038","title":"LSSF: Safety Alignment for Large Language Models through Low-Rank Safety Subspace Fusion","created":"2026-01-19","cited_by_count":8,"area":"alignment"},{"id":"2502.20757","title":"The Rise of Darkness: Safety-Utility Trade-Offs in Role-Playing Dialogue Agents","created":"2025-02-28","cited_by_count":7,"area":"agents"},{"id":"2507.21141","title":"The Geometry of Harmfulness in LLMs through Subconcept Probing","created":"2025-07-23","cited_by_count":7,"area":null},{"id":"2509.06795","title":"Anchoring Refusal Direction: Mitigating Safety Risks in Tuning via Projection Constraint","created":"2025-09-08","cited_by_count":7,"area":null},{"id":"2601.10141","title":"Understanding and Preserving Safety in Fine-Tuned LLMs","created":"2026-01-15","cited_by_count":7,"area":"safety"},{"id":"2411.06824","title":"Combining Domain and Alignment Vectors to Achieve Better Knowledge-Safety Trade-offs in LLMs","created":"2024-11-11","cited_by_count":6,"area":null},{"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":"2504.20106","title":"Adaptive Helpfulness-Harmlessness Alignment with Preference Vectors","created":"2025-04-27","cited_by_count":6,"area":"alignment"},{"id":"2505.14185","title":"Safety Subspaces are Not Linearly Distinct: A Fine-Tuning Case Study","created":"2025-05-20","cited_by_count":6,"area":"alignment"},{"id":"2601.10589","title":"Be Your Own Red Teamer: Safety Alignment via Self-Play and Reflective Experience Replay","created":"2026-01-15","cited_by_count":6,"area":"alignment"},{"id":"2602.22554","title":"Multilingual Safety Alignment Via Sparse Weight Editing","created":"2026-02-26","cited_by_count":6,"area":"alignment"},{"id":"2505.17519","title":"Chain-of-Lure: A Universal Jailbreak Attack Framework using Unconstrained Synthetic Narratives","created":"2025-05-23","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":"2510.11570","title":"Strong but Brittle: Simple Attacks Subvert Reasoning-based Safety Guardrails","created":"2025-10-13","cited_by_count":5,"area":"lr"},{"id":"2601.08089","title":"Q-realign: Piggybacking Realignment on Quantization for Safe and Efficient LLM Deployment","created":"2026-01-13","cited_by_count":5,"area":"safety"},{"id":"2505.00649","title":"Investigating Task Arithmetic for Zero-Shot Information Retrieval","created":"2025-05-01","cited_by_count":4,"area":null},{"id":"2601.04740","title":"StealthGraph: Exposing Domain-Specific Risks in LLMs through Knowledge-Graph-Guided Harmful Prompt Generation","created":"2026-01-08","cited_by_count":4,"area":"safety"},{"id":"2601.17642","title":"Health-ORSC-Bench: A Benchmark for Measuring Over-Refusal and Safety Completion in Health Context","created":"2026-01-25","cited_by_count":4,"area":"alignment"},{"id":"2602.05228","title":"Surgery: Mitigating Harmful Fine-Tuning for Large Language Models via Attention Sink","created":"2026-02-05","cited_by_count":4,"area":"alignment"},{"id":"2602.08741","title":"Large Language Lobotomy: Jailbreaking Mixture-of-Experts via Expert Silencing","created":"2026-02-09","cited_by_count":4,"area":"safety"},{"id":"2501.02373","title":"BADTV: Unveiling Backdoor Threats in Third-Party Task Vectors","created":"2025-01-04","cited_by_count":3,"area":"safety"},{"id":"2510.05310","title":"RAG Makes Guardrails Unsafe? Investigating Robustness of Guardrails under RAG-style Contexts","created":"2025-10-06","cited_by_count":3,"area":"safety"},{"id":"2601.03699","title":"RedBench: A Universal Dataset for Comprehensive Red Teaming of Large Language Models","created":"2026-01-07","cited_by_count":3,"area":"safety"},{"id":"2602.07013","title":"Steering to Say No: Configurable Refusal via Activation Steering in Vision Language Models","created":"2026-01-31","cited_by_count":3,"area":"mllm"},{"id":"2605.04446","title":"Misrouter: Exploiting Routing Mechanisms for Input-Only Attacks on Mixture-of-Experts LLMs","created":"2026-05-06","cited_by_count":3,"area":"safety"},{"id":"2607.12624","title":"PVDetector: Detecting Prompt Injection Attacks on Purpose-Specific LLM Agents through Policy-Violation Concept Analysis","created":"2026-07-14","cited_by_count":3,"area":"agents"},{"id":"2507.08898","title":"SEALGuard: Safeguarding the Multilingual Conversations in Southeast Asian Languages for LLM Software Systems","created":"2025-07-11","cited_by_count":2,"area":"alignment"},{"id":"2508.12622","title":"Consiglieres in the Shadow: Understanding the Use of Uncensored Large Language Models in Cybercrimes","created":"2025-08-18","cited_by_count":2,"area":"safety"},{"id":"2508.16325","title":"ConceptGuard: Neuro-Symbolic Safety Guardrails via Sparse Interpretable Jailbreak Concepts","created":"2025-08-22","cited_by_count":2,"area":"safety"},{"id":"2511.10712","title":"Do Not Merge My Model! Safeguarding Open-Source LLMs Against Unauthorized Model Merging","created":"2025-11-13","cited_by_count":2,"area":null},{"id":"2602.13071","title":"Bus-Conditioned Zero-Shot Trajectory Generation via Task Arithmetic","created":"2026-02-13","cited_by_count":2,"area":null},{"id":"2602.16835","title":"NeST: Neuron Selective Tuning for LLM Safety","created":"2026-02-18","cited_by_count":2,"area":"alignment"},{"id":"2605.24817","title":"RouteScan: A Non-Intrusive Approach to Auditing MoE LLMs Safety via Expert Routing Telemetry","created":"2026-05-24","cited_by_count":2,"area":"moe"},{"id":"2606.07535","title":"Multilingual Refusal Alignment for Safer Large Language Models","created":"2026-04-24","cited_by_count":2,"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":"2509.16870","title":"DecipherGuard: Understanding and Deciphering Jailbreak Prompts for a Safer Deployment of Intelligent Software Systems","created":"2025-09-21","cited_by_count":1,"area":"safety"},{"id":"2510.11268","title":"Exploring and Leveraging Class Vectors for Classifier Editing","created":"2025-10-13","cited_by_count":1,"area":null},{"id":"2604.15780","title":"Pruning Unsafe Tickets: A Resource-Efficient Framework for Safer and More Robust LLMs","created":"2026-04-17","cited_by_count":1,"area":"safety"},{"id":"2605.28553","title":"Refusal Before Decoding: Detecting and Exploiting Refusal Signals in Intermediate LLM Activations","created":"2026-05-27","cited_by_count":1,"area":"mechinterp"},{"id":"2606.20668","title":"BELLS-O: Evaluating the Operational Trade-offs of LLM Supervision Systems","created":"2026-06-12","cited_by_count":1,"area":"safety"},{"id":"2607.27081","title":"On-Policy Distillation for LLM Safety: A Routing Approach to Template-Robust Realignment","created":"2026-07-29","cited_by_count":1,"area":"safety"},{"id":"2502.12189","title":"Self-supervised Attribute-aware Dynamic Preference Ranking Alignment","created":"2025-02-15","cited_by_count":0,"area":"alignment"},{"id":"2511.12487","title":"ToxSearch: Evolving Prompts for Toxicity Search in Large Language Models","created":"2025-11-16","cited_by_count":0,"area":"alignment"},{"id":"2601.04199","title":"The Forgotten Shield: Safety Grafting in Parameter-Space for Medical MLLMs","created":"2025-12-05","cited_by_count":0,"area":"mllm"},{"id":"2603.22214","title":"Evaluating the Reliability and Fidelity of Automated Judgment Systems of Large Language Models","created":"2026-03-23","cited_by_count":0,"area":null},{"id":"2604.09265","title":"EthicMind: A Risk-Aware Framework for Ethical-Emotional Alignment in Multi-Turn Dialogue","created":"2026-04-10","cited_by_count":0,"area":null},{"id":"2605.04572","title":"From Parameter Dynamics to Risk Scoring : Quantifying Sample-Level Safety Degradation in LLM Fine-tuning","created":"2026-05-06","cited_by_count":0,"area":"alignment"},{"id":"2605.13411","title":"Model-Agnostic Lifelong LLM Safety via Externalized Attack-Defense Co-Evolution","created":"2026-05-13","cited_by_count":0,"area":"safety"},{"id":"2605.30640","title":"CSULoRA: Closest Safe Update Low-Rank Adaptation","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":"2606.09697","title":"PsychoSafe: Eliciting Psychologically-Informed Refusals in Large Language Models","created":"2026-06-08","cited_by_count":0,"area":null},{"id":"2606.27944","title":"It Lied to a Doctor to Buy Poison Ingredients: Quantifying Real-World Misuse of Phone-use Agents","created":"2026-06-26","cited_by_count":0,"area":null},{"id":"2606.28525","title":"A Gravitational Interpretation of Safety Reversion under Fine-Tuning","created":"2026-06-26","cited_by_count":0,"area":"safety"},{"id":"2607.11475","title":"HyperSafe: Inference-Time Safety Recovery for Fine-Tuned Language Models","created":"2026-07-13","cited_by_count":0,"area":"alignment"},{"id":"2607.16242","title":"TRACE: Trajectory-Based Safety Patch Learning for LLM Post-Training Realignment","created":"2026-06-26","cited_by_count":0,"area":"alignment"},{"id":"2608.00042","title":"Trustworthiness Costs of Domain Adaptation in Small Language Models:A Cross-Architecture Empirical Study","created":"2026-07-23","cited_by_count":0,"area":null}],"next_cursor":"100"}