{"total":209,"rows":[{"id":"2307.15043","title":"Universal and Transferable Adversarial Attacks on Aligned Language Models","created":"2023-07-27","cited_by_count":3104,"area":"alignment"},{"id":"2310.03693","title":"Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!","created":"2023-10-05","cited_by_count":1154,"area":"alignment"},{"id":"2310.04451","title":"AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models","created":"2023-10-03","cited_by_count":785,"area":"alignment"},{"id":"2307.15217","title":"Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback","created":"2023-07-27","cited_by_count":759,"area":"alignment"},{"id":"2406.05946","title":"Safety Alignment Should Be Made More Than Just a Few Tokens Deep","created":"2024-06-10","cited_by_count":472,"area":"alignment"},{"id":"2310.06387","title":"Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations","created":"2023-10-10","cited_by_count":431,"area":"alignment"},{"id":"2308.06463","title":"GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher","created":"2023-08-12","cited_by_count":423,"area":"alignment"},{"id":"2405.20947","title":"OR-Bench: An Over-Refusal Benchmark for Large Language Models","created":"2024-05-31","cited_by_count":244,"area":"alignment"},{"id":"2402.05162","title":"Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications","created":"2024-02-07","cited_by_count":231,"area":"alignment"},{"id":"2309.14348","title":"Defending Against Alignment-Breaking Attacks via Robustly Aligned LLM","created":"2023-09-18","cited_by_count":192,"area":"alignment"},{"id":"2401.01967","title":"A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity","created":"2024-01-03","cited_by_count":185,"area":"alignment"},{"id":"2309.10105","title":"Understanding Catastrophic Forgetting in Language Models via Implicit Inference","created":"2023-09-18","cited_by_count":144,"area":"alignment"},{"id":"2403.17710","title":"Optimization-based Prompt Injection Attack to LLM-as-a-Judge","created":"2024-03-26","cited_by_count":143,"area":"alignment"},{"id":"2312.08358","title":"Distributional Preference Learning: Understanding and Accounting for Hidden Context in RLHF","created":"2023-12-13","cited_by_count":133,"area":"alignment"},{"id":"2311.14455","title":"Universal Jailbreak Backdoors from Poisoned Human Feedback","created":"2023-11-24","cited_by_count":128,"area":"alignment"},{"id":"2402.01109","title":"Vaccine: Perturbation-aware Alignment for Large Language Models against Harmful Fine-tuning Attack","created":"2024-02-02","cited_by_count":115,"area":"alignment"},{"id":"2310.15140","title":"AutoDAN: Interpretable Gradient-Based Adversarial Attacks on Large Language Models","created":"2023-10-23","cited_by_count":111,"area":"alignment"},{"id":"2402.19085","title":"Controllable Preference Optimization: Toward Controllable Multi-Objective Alignment","created":"2024-02-29","cited_by_count":107,"area":"alignment"},{"id":"2406.05644","title":"How Alignment and Jailbreak Work: Explain LLM Safety through Intermediate Hidden States","created":"2024-06-09","cited_by_count":98,"area":"alignment"},{"id":"2401.13136","title":"The Language Barrier: Dissecting Safety Challenges of LLMs in Multilingual Contexts","created":"2024-01-23","cited_by_count":94,"area":"alignment"},{"id":"2405.18641","title":"Lisa: Lazy Safety Alignment for Large Language Models against Harmful Fine-tuning Attack","created":"2024-05-28","cited_by_count":92,"area":"alignment"},{"id":"2404.08555","title":"RLHF Deciphered: A Critical Analysis of Reinforcement Learning from Human Feedback for LLMs","created":"2024-04-12","cited_by_count":90,"area":"alignment"},{"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":"2404.12038","title":"Uncovering Safety Risks of Large Language Models through Concept Activation Vector","created":"2024-04-18","cited_by_count":67,"area":"alignment"},{"id":"2404.16369","title":"Don't Say No: Jailbreaking LLM by Suppressing Refusal","created":"2024-04-25","cited_by_count":56,"area":"alignment"},{"id":"2402.06147","title":"DeAL: Decoding-time Alignment for Large Language Models","created":"2024-02-05","cited_by_count":55,"area":"alignment"},{"id":"2403.02475","title":"Enhancing LLM Safety via Constrained Direct Preference Optimization","created":"2024-03-04","cited_by_count":49,"area":"alignment"},{"id":"2405.17374","title":"Navigating the Safety Landscape: Measuring Risks in Finetuning Large Language Models","created":"2024-05-27","cited_by_count":46,"area":"alignment"},{"id":"2407.02273","title":"Language Model Alignment in Multilingual Trolley Problems","created":"2024-07-02","cited_by_count":46,"area":"alignment"},{"id":"2405.19358","title":"Robustifying Safety-Aligned Large Language Models through Clean Data Curation","created":"2024-05-24","cited_by_count":44,"area":"alignment"},{"id":"2406.14563","title":"Model Merging and Safety Alignment: One Bad Model Spoils the Bunch","created":"2024-06-20","cited_by_count":43,"area":"alignment"},{"id":"2311.00168","title":"The Alignment Ceiling: Objective Mismatch in Reinforcement Learning from Human Feedback","created":"2023-10-31","cited_by_count":42,"area":"alignment"},{"id":"2506.07468","title":"Chasing Moving Targets with Online Self-Play Reinforcement Learning for Safer Language Models","created":"2025-06-09","cited_by_count":39,"area":"alignment"},{"id":"2504.09466","title":"AdaSteer: Your Aligned LLM is Inherently an Adaptive Jailbreak Defender","created":"2025-04-13","cited_by_count":38,"area":"alignment"},{"id":"2311.06899","title":"Flames: Benchmarking Value Alignment of LLMs in Chinese","created":"2023-11-12","cited_by_count":37,"area":"alignment"},{"id":"2405.13967","title":"Model Editing as a Robust and Denoised variant of DPO: A Case Study on Toxicity","created":"2024-05-22","cited_by_count":35,"area":"alignment"},{"id":"2409.14586","title":"Backtracking Improves Generation Safety","created":"2024-09-22","cited_by_count":32,"area":"alignment"},{"id":"2406.12935","title":"ChatBug: A Common Vulnerability of Aligned LLMs Induced by Chat Templates","created":"2024-06-17","cited_by_count":30,"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":"2405.20653","title":"Mind the Inconspicuous: Revealing the Hidden Weakness in Aligned LLMs' Refusal Boundaries","created":"2024-05-31","cited_by_count":28,"area":"alignment"},{"id":"2310.01581","title":"On the Safety of Open-Sourced Large Language Models: Does Alignment Really Prevent Them From Being Misused?","created":"2023-10-02","cited_by_count":26,"area":"alignment"},{"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":"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":"2408.01420","title":"Mission Impossible: A Statistical Perspective on Jailbreaking LLMs","created":"2024-08-02","cited_by_count":24,"area":"alignment"},{"id":"2406.08725","title":"RL-JACK: Reinforcement Learning-powered Black-box Jailbreaking Attack against LLMs","created":"2024-06-13","cited_by_count":23,"area":"alignment"},{"id":"2409.11564","title":"Preference Tuning with Human Feedback on Language, Speech, and Vision Tasks: A Survey","created":"2024-09-17","cited_by_count":23,"area":"alignment"},{"id":"2407.07342","title":"Multilingual Blending: LLM Safety Alignment Evaluation with Language Mixture","created":"2024-07-10","cited_by_count":22,"area":"alignment"},{"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":"2602.01539","title":"MAGIC: A Co-Evolving Attacker-Defender Adversarial Game for Robust LLM Safety","created":"2026-02-02","cited_by_count":21,"area":"alignment"},{"id":"2502.09723","title":"QueryAttack: Jailbreaking Aligned Large Language Models Using Structured Non-natural Query Language","created":"2025-02-13","cited_by_count":18,"area":"alignment"},{"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":"2409.02795","title":"Towards a Unified View of Preference Learning for Large Language Models: A Survey","created":"2024-09-04","cited_by_count":17,"area":"alignment"},{"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":"2502.04322","title":"Speak Easy: Eliciting Harmful Jailbreaks from LLMs with Simple Interactions","created":"2025-02-06","cited_by_count":15,"area":"alignment"},{"id":"2505.20259","title":"Lifelong Safety Alignment for Language Models","created":"2025-05-26","cited_by_count":15,"area":"alignment"},{"id":"2507.14987","title":"AlphaAlign: Incentivizing Safety Alignment with Extremely Simplified Reinforcement Learning","created":"2025-07-20","cited_by_count":15,"area":"alignment"},{"id":"2410.10862","title":"Superficial Safety Alignment Hypothesis","created":"2024-10-07","cited_by_count":14,"area":"alignment"},{"id":"2407.16637","title":"Course-Correction: Safety Alignment Using Synthetic Preferences","created":"2024-07-23","cited_by_count":13,"area":"alignment"},{"id":"2412.01253","title":"Yi-Lightning Technical Report","created":"2024-12-02","cited_by_count":12,"area":"alignment"},{"id":"2505.17598","title":"One Model Transfer to All: On Robust Jailbreak Prompts Generation against LLMs","created":"2025-05-23","cited_by_count":12,"area":"alignment"},{"id":"2505.21605","title":"SoSBench: Benchmarking Safety Alignment on Six Scientific Domains","created":"2025-05-27","cited_by_count":12,"area":"alignment"},{"id":"2312.07876","title":"Causality Analysis for Evaluating the Security of Large Language Models","created":"2023-12-13","cited_by_count":11,"area":"alignment"},{"id":"2404.00486","title":"Dialectical Alignment: Resolving the Tension of 3H and Security Threats of LLMs","created":"2024-03-30","cited_by_count":11,"area":"alignment"},{"id":"2405.17382","title":"ReMoDetect: Reward Models Recognize Aligned LLM's Generations","created":"2024-05-27","cited_by_count":11,"area":"alignment"},{"id":"2406.14393","title":"Jailbreaking as a Reward Misspecification Problem","created":"2024-06-20","cited_by_count":11,"area":"alignment"},{"id":"2411.12762","title":"Playing Language Game with LLMs Leads to Jailbreaking","created":"2024-11-16","cited_by_count":11,"area":"alignment"},{"id":"2501.16727","title":"xJailbreak: Representation Space Guided Reinforcement Learning for Interpretable LLM Jailbreaking","created":"2025-01-28","cited_by_count":11,"area":"alignment"},{"id":"2512.11391","title":"Mitigating the Safety Alignment Tax with Null-Space Constrained Policy Optimization","created":"2025-12-12","cited_by_count":11,"area":"alignment"},{"id":"2602.02495","title":"Reward-free Alignment for Conflicting Objectives","created":"2026-02-02","cited_by_count":11,"area":"alignment"},{"id":"2405.11647","title":"Hummer: Towards Limited Competitive Preference Dataset","created":"2024-05-19","cited_by_count":10,"area":"alignment"},{"id":"2408.14866","title":"Advancing Adversarial Suffix Transfer Learning on Aligned Large Language Models","created":"2024-08-27","cited_by_count":10,"area":"alignment"},{"id":"2502.15435","title":"Single-pass Detection of Jailbreaking Input in Large Language Models","created":"2025-02-21","cited_by_count":10,"area":"alignment"},{"id":"2509.21528","title":"Preemptive Detection and Steering of LLM Misalignment via Latent Reachability","created":"2025-09-25","cited_by_count":10,"area":"alignment"},{"id":"2603.01246","title":"Defensive Refusal Bias: How Safety Alignment Fails Cyber Defenders","created":"2026-03-01","cited_by_count":10,"area":"alignment"},{"id":"2406.18118","title":"SafeAligner: Safety Alignment against Jailbreak Attacks via Response Disparity Guidance","created":"2024-06-26","cited_by_count":9,"area":"alignment"},{"id":"2408.05094","title":"Unlocking Decoding-time Controllability: Gradient-Free Multi-Objective Alignment with Contrastive Prompts","created":"2024-08-09","cited_by_count":9,"area":"alignment"},{"id":"2505.17072","title":"Safety Alignment Can Be Not Superficial With Explicit Safety Signals","created":"2025-05-19","cited_by_count":9,"area":"alignment"},{"id":"2512.13655","title":"Comparative Analysis of LLM Abliteration Methods: A Cross-Architecture Evaluation","created":"2025-12-15","cited_by_count":9,"area":"alignment"},{"id":"2604.19018","title":"Local Linearity of LLMs Enables Activation Steering via Model-Based Linear Optimal Control","created":"2026-04-21","cited_by_count":9,"area":"alignment"},{"id":"2311.07723","title":"Generalization Analogies: A Testbed for Generalizing AI Oversight to Hard-To-Measure Domains","created":"2023-11-13","cited_by_count":8,"area":"alignment"},{"id":"2406.15599","title":"Pareto-Optimal Learning from Preferences with Hidden Context","created":"2024-06-21","cited_by_count":8,"area":"alignment"},{"id":"2406.16743","title":"Safety Alignment of Large Language Models via Contrasting Safe and Harmful Distributions","created":"2024-06-24","cited_by_count":8,"area":"alignment"},{"id":"2410.12999","title":"POROver: Improving Safety and Reducing Overrefusal in Large Language Models with Overgeneration and Preference Optimization","created":"2024-10-16","cited_by_count":8,"area":"alignment"},{"id":"2503.00069","title":"Societal Alignment Frameworks Can Improve LLM Alignment","created":"2025-02-27","cited_by_count":8,"area":"alignment"},{"id":"2506.13901","title":"Alignment Quality Index (AQI) : Beyond Refusals: AQI as an Intrinsic Alignment Diagnostic via Latent Geometry, Cluster Divergence, and Layer wise Pooled Representations","created":"2025-06-16","cited_by_count":8,"area":"alignment"},{"id":"2507.11500","title":"ARMOR: Aligning Secure and Safe Large Language Models via Meticulous Reasoning","created":"2025-07-14","cited_by_count":8,"area":"alignment"},{"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":"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"},{"id":"2407.06542","title":"LIONs: An Empirically Optimized Approach to Align Language Models","created":"2024-07-09","cited_by_count":7,"area":"alignment"},{"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":"2508.14904","title":"Efficient Switchable Safety Control in LLMs via Magic-Token-Guided Co-Training","created":"2025-08-12","cited_by_count":7,"area":"alignment"},{"id":"2510.05052","title":"Proactive defense against LLM Jailbreak","created":"2025-10-06","cited_by_count":7,"area":"alignment"},{"id":"2601.03868","title":"What Matters For Safety Alignment?","created":"2026-01-07","cited_by_count":7,"area":"alignment"},{"id":"2603.11388","title":"Deactivating Refusal Triggers: Understanding and Mitigating Overrefusal in Safety Alignment","created":"2026-03-12","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":"2502.00657","title":"LLM Safety Alignment is Divergence Estimation in Disguise","created":"2025-02-02","cited_by_count":6,"area":"alignment"},{"id":"2502.01925","title":"PANDAS: Improving Many-shot Jailbreaking via Positive Affirmation, Negative Demonstration, and Adaptive Sampling","created":"2025-02-04","cited_by_count":6,"area":"alignment"},{"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":"2502.13946","title":"Why Safeguarded Ships Run Aground? Aligned Large Language Models' Safety Mechanisms Tend to Be Anchored in The Template Region","created":"2025-02-19","cited_by_count":6,"area":"alignment"}],"next_cursor":"100"}