{"total":114,"rows":[{"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":"2403.13787","title":"RewardBench: Evaluating Reward Models for Language Modeling","created":"2024-03-20","cited_by_count":499,"area":"alignment"},{"id":"2310.02949","title":"Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models","created":"2023-10-04","cited_by_count":297,"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":"2406.14598","title":"SORRY-Bench: Systematically Evaluating Large Language Model Safety Refusal","created":"2024-06-20","cited_by_count":240,"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":"2410.16184","title":"RM-Bench: Benchmarking Reward Models of Language Models with Subtlety and Style","created":"2024-10-21","cited_by_count":176,"area":"alignment"},{"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":"2507.18624","title":"Checklists Are Better Than Reward Models For Aligning Language Models","created":"2025-07-24","cited_by_count":128,"area":"alignment"},{"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":"2411.01111","title":"Rule Based Rewards for Language Model Safety","created":"2024-11-02","cited_by_count":108,"area":"alignment"},{"id":"2405.20850","title":"Improving Reward Models with Synthetic Critiques","created":"2024-05-31","cited_by_count":44,"area":"alignment"},{"id":"2409.20370","title":"The Perfect Blend: Redefining RLHF with Mixture of Judges","created":"2024-09-30","cited_by_count":43,"area":"alignment"},{"id":"2503.03710","title":"Improving LLM Safety Alignment with Dual-Objective Optimization","created":"2025-03-05","cited_by_count":43,"area":"alignment"},{"id":"2501.12895","title":"Test-Time Preference Optimization: On-the-Fly Alignment via Iterative Textual Feedback","created":"2025-01-22","cited_by_count":41,"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":"2505.20065","title":"SafeDPO: A Simple Approach to Direct Preference Optimization with Enhanced Safety","created":"2025-05-26","cited_by_count":37,"area":"alignment"},{"id":"2409.00598","title":"Automatic Pseudo-Harmful Prompt Generation for Evaluating False Refusals in Large Language Models","created":"2024-09-01","cited_by_count":35,"area":"alignment"},{"id":"2409.15268","title":"Style Outweighs Substance: Failure Modes of LLM Judges in Alignment Benchmarking","created":"2024-09-23","cited_by_count":35,"area":"alignment"},{"id":"2407.13833","title":"Phi-3 Safety Post-Training: Aligning Language Models with a \"Break-Fix\" Cycle","created":"2024-07-18","cited_by_count":32,"area":"alignment"},{"id":"2401.16332","title":"Tradeoffs Between Alignment and Helpfulness in Language Models with Steering Methods","created":"2024-01-29","cited_by_count":28,"area":"alignment"},{"id":"2506.03637","title":"RewardAnything: Generalizable Principle-Following Reward Models","created":"2025-06-04","cited_by_count":28,"area":"alignment"},{"id":"2410.02197","title":"Beyond Bradley-Terry Models: A General Preference Model for Language Model Alignment","created":"2024-10-03","cited_by_count":27,"area":"alignment"},{"id":"2507.04250","title":"Just Enough Shifts: Mitigating Over-Refusal in Aligned Language Models with Targeted Representation Fine-Tuning","created":"2025-07-06","cited_by_count":24,"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":"2501.14940","title":"CASE-Bench: Context-Aware SafEty Benchmark for Large Language Models","created":"2025-01-24","cited_by_count":20,"area":"alignment"},{"id":"2502.09990","title":"X-Boundary: Establishing Exact Safety Boundary to Shield LLMs from Multi-Turn Jailbreaks without Compromising Usability","created":"2025-02-14","cited_by_count":20,"area":"alignment"},{"id":"2503.18991","title":"Inverse Reinforcement Learning with Dynamic Reward Scaling for LLM Alignment","created":"2025-03-23","cited_by_count":20,"area":"alignment"},{"id":"2509.01909","title":"Oyster-I: Beyond Refusal -- Constructive Safety Alignment for Responsible Language Models","created":"2025-09-02","cited_by_count":19,"area":"alignment"},{"id":"2503.02574","title":"LLM-Safety Evaluations Lack Robustness","created":"2025-03-04","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":"2406.10288","title":"Do as I do (Safely): Mitigating Task-Specific Fine-tuning Risks in Large Language Models","created":"2024-06-12","cited_by_count":17,"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":"2410.15483","title":"Understanding Forgetting in LLM Supervised Fine-Tuning and Preference Learning -- A Convex Optimization Perspective","created":"2024-10-20","cited_by_count":12,"area":"alignment"},{"id":"2412.16974","title":"Cannot or Should Not? Automatic Analysis of Refusal Composition in IFT/RLHF Datasets and Refusal Behavior of Black-Box LLMs","created":"2024-12-22","cited_by_count":11,"area":"alignment"},{"id":"2506.21463","title":"Aligning Spoken Dialogue Models from User Interactions","created":"2025-06-26","cited_by_count":11,"area":"alignment"},{"id":"2512.10150","title":"Unforgotten Safety: Preserving Safety Alignment of Large Language Models with Continual Learning","created":"2025-12-10","cited_by_count":10,"area":"alignment"},{"id":"2411.06835","title":"HarmLevelBench: Evaluating Harm-Level Compliance and the Impact of Quantization on Model Alignment","created":"2024-11-11","cited_by_count":9,"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":"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":"2512.20806","title":"Safety Alignment of LMs via Non-cooperative Games","created":"2025-12-23","cited_by_count":8,"area":"alignment"},{"id":"2410.23726","title":"Towards Reliable Alignment: Uncertainty-aware RLHF","created":"2024-10-31","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":"2505.00038","title":"HyPerAlign: Interpretable Personalized LLM Alignment via Hypothesis Generation","created":"2025-04-29","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":"2601.03868","title":"What Matters For Safety Alignment?","created":"2026-01-07","cited_by_count":7,"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":"2505.23473","title":"EVOREFUSE: Evolutionary Prompt Optimization for Evaluation and Mitigation of LLM Over-Refusal to Pseudo-Malicious Instructions","created":"2025-05-29","cited_by_count":6,"area":"alignment"},{"id":"2510.04320","title":"Read the Scene, Not the Script: Outcome-Aware Safety for LLMs","created":"2025-10-05","cited_by_count":6,"area":"alignment"},{"id":"2602.07892","title":"Safety Alignment as Continual Learning: Mitigating the Alignment Tax via Orthogonal Gradient Projection","created":"2026-02-08","cited_by_count":6,"area":"alignment"},{"id":"2410.19933","title":"Enhancing Safety in Reinforcement Learning with Human Feedback via Rectified Policy Optimization","created":"2024-10-25","cited_by_count":5,"area":"alignment"},{"id":"2508.09473","title":"NeuronTune: Fine-Grained Neuron Modulation for Balanced Safety-Utility Alignment in LLMs","created":"2025-08-13","cited_by_count":4,"area":"alignment"},{"id":"2510.13698","title":"Attention Misses Visual Risk: Risk-Adaptive Steering for Multimodal Safety Alignment","created":"2025-10-15","cited_by_count":4,"area":"alignment"},{"id":"2604.18901","title":"Harmful Intent as a Geometrically Recoverable Feature of LLM Residual Streams","created":"2026-04-20","cited_by_count":4,"area":"alignment"},{"id":"2506.04462","title":"Watermarking Degrades Alignment in Language Models: Analysis and Mitigation","created":"2025-06-04","cited_by_count":3,"area":"alignment"},{"id":"2510.18081","title":"Any-Depth Alignment: Unlocking Innate Safety Alignment of LLMs to Any-Depth","created":"2025-10-20","cited_by_count":3,"area":"alignment"},{"id":"2511.12464","title":"Probing Preference Representations: A Multi-Dimensional Evaluation and Analysis Method for Reward Models","created":"2025-11-16","cited_by_count":3,"area":"alignment"},{"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":"2604.27093","title":"Useless but Safe? Benchmarking Utility Recovery with User Intent Clarification in Multi-Turn Conversations","created":"2026-04-29","cited_by_count":3,"area":"alignment"},{"id":"2512.01037","title":"How Semantically Stable Are LLM Refusals? Measuring Confusion in Local Safety Boundaries","created":"2025-11-30","cited_by_count":2,"area":"alignment"},{"id":"2602.16977","title":"Fail-Closed Alignment for Large Language Models","created":"2026-02-19","cited_by_count":2,"area":"alignment"},{"id":"2606.02995","title":"Patcher: Post-Hoc Patching of Backdoored Large Language Models","created":"2026-06-02","cited_by_count":2,"area":"alignment"},{"id":"2606.06797","title":"Korean Culture into LLM Alignment: Toward Cultural Coherence","created":"2026-06-05","cited_by_count":2,"area":"alignment"},{"id":"2511.00509","title":"Reimagining Safety Alignment with An Image","created":"2025-11-01","cited_by_count":1,"area":"alignment"},{"id":"2601.03300","title":"TRYLOCK: Defense-in-Depth Against LLM Jailbreaks via Layered Preference and Representation Engineering","created":"2026-01-06","cited_by_count":1,"area":"alignment"},{"id":"2603.13359","title":"From Refusal Tokens to Refusal Control: Discovering and Steering Category-Specific Refusal Directions","created":"2026-03-09","cited_by_count":1,"area":"alignment"},{"id":"2603.22829","title":"Improving Safety Alignment via Balanced Direct Preference Optimization","created":"2026-03-24","cited_by_count":1,"area":"alignment"},{"id":"2604.00979","title":"Dual Optimal: Make Your LLM Peer-like with Dignity","created":"2026-04-01","cited_by_count":1,"area":"alignment"},{"id":"2604.02669","title":"Redirected, Not Removed: Task-Dependent Stereotyping Reveals the Limits of LLM Alignments","created":"2026-04-03","cited_by_count":1,"area":"alignment"},{"id":"2604.09189","title":"Do LLMs Follow Their Own Rules? A Reflexive Audit of Self-Stated Safety Policies","created":"2026-04-10","cited_by_count":1,"area":"alignment"},{"id":"2604.17299","title":"Cat-DPO: Category-Adaptive Safety Alignment","created":"2026-04-19","cited_by_count":1,"area":"alignment"},{"id":"2605.01899","title":"Disentangling Intent from Role: Adversarial Self-Play for Persona-Invariant Safety Alignment","created":"2026-05-03","cited_by_count":1,"area":"alignment"},{"id":"2605.24154","title":"Palette: A Modular, Controllable, and Efficient Framework for On-demand Authorized Safety Alignment Relaxation in LLMs","created":"2026-05-22","cited_by_count":1,"area":"alignment"},{"id":"2606.19168","title":"Beyond Safe Data: Pretraining-Stage Alignment with Regular Safety Reflection","created":"2026-06-17","cited_by_count":1,"area":"alignment"},{"id":"2609.22219","title":"Knowing, and Saying It Only When Asked: LLM Endognostics and the Schizognosis of Minerva-7B","created":"2026-09-02","cited_by_count":1,"area":"alignment"},{"id":"2501.00911","title":"Aligning LLMs with Domain Invariant Reward Models","created":"2025-01-01","cited_by_count":0,"area":"alignment"},{"id":"2508.09937","title":"A Comprehensive Evaluation framework of Alignment Techniques for LLMs","created":"2025-08-13","cited_by_count":0,"area":"alignment"},{"id":"2509.12221","title":"MEUV: Achieving Fine-Grained Capability Activation in Large Language Models via Mutually Exclusive Unlock Vectors","created":"2025-09-04","cited_by_count":0,"area":"alignment"},{"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":"2604.06833","title":"FedDetox: Robust Federated SLM Alignment via On-Device Data Sanitization","created":"2026-04-08","cited_by_count":0,"area":"alignment"},{"id":"2604.07754","title":"The Art of (Mis)alignment: How Fine-Tuning Methods Effectively Misalign and Realign LLMs in Post-Training","created":"2026-04-09","cited_by_count":0,"area":"alignment"},{"id":"2605.04992","title":"You Snooze, You Lose: Automatic Safety Alignment Restoration through Neural Weight Translation","created":"2026-05-06","cited_by_count":0,"area":"alignment"},{"id":"2605.07883","title":"Beyond \"I cannot fulfill this request\": Alleviating Rigid Rejection in LLMs via Label Enhancement","created":"2026-05-08","cited_by_count":0,"area":"alignment"},{"id":"2605.08277","title":"Mitigating Many-shot Jailbreak Attacks with One Single Demonstration","created":"2026-05-08","cited_by_count":0,"area":"alignment"},{"id":"2605.10998","title":"Few-Shot Truly Benign DPO Attack for Jailbreaking LLMs","created":"2026-05-09","cited_by_count":0,"area":"alignment"},{"id":"2605.11712","title":"Toward Stable Value Alignment: Introducing Independent Modules for Consistent Value Guidance","created":"2026-05-12","cited_by_count":0,"area":"alignment"},{"id":"2605.20654","title":"REFLECTOR: Internalizing Step-wise Reflection against Indirect Jailbreak","created":"2026-05-20","cited_by_count":0,"area":"alignment"},{"id":"2605.26315","title":"Curriculum Learning for Safety Alignment","created":"2026-05-25","cited_by_count":0,"area":"alignment"},{"id":"2605.30487","title":"Configurable Reward Model for Balanced Safety Alignment","created":"2026-05-28","cited_by_count":0,"area":"alignment"},{"id":"2606.04778","title":"Inference-Time Vulnerability Beyond Shallow Safety: Alignment Along Generation Trajectories","created":"2026-06-03","cited_by_count":0,"area":"alignment"},{"id":"2606.09635","title":"Gradient-Guided Reward Optimization for Inference-time Alignment","created":"2026-06-08","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":"2606.29581","title":"The Joint Effect of Quantization and Sampling Temperature on LLM Safety Alignment: A Factorial Analysis","created":"2026-06-28","cited_by_count":0,"area":"alignment"},{"id":"2607.00572","title":"HARC: Coupling Harmfulness and Refusal Directions for Robust Safety Alignment","created":"2026-07-01","cited_by_count":0,"area":"alignment"},{"id":"2607.02914","title":"Oyster-II: Reinforcement Learning for Constructive Safety Alignment in Large Language Models","created":"2026-07-03","cited_by_count":0,"area":"alignment"},{"id":"2607.14147","title":"Breaking Refusal in the First Half: A Mechanistic Study of the Prefill Jailbreak","created":"2026-07-14","cited_by_count":0,"area":"alignment"},{"id":"2607.14480","title":"Lower-Resource, Higher Scores: Language Bias in LLM Evaluators","created":"2026-07-16","cited_by_count":0,"area":"alignment"},{"id":"2607.21151","title":"V-DEAL: Diagnosing Video Safety De-Calibration as an Understanding-Refusal Coupling Failure","created":"2026-07-23","cited_by_count":0,"area":"alignment"},{"id":"2607.22305","title":"A Roadmap to Impactful Pluralistic Alignment Research","created":"2026-07-24","cited_by_count":0,"area":"alignment"}],"next_cursor":"100"}