{"total":92,"rows":[{"id":"2406.00045","title":"Personalized Steering of Large Language Models: Versatile Steering Vectors Through Bi-directional Preference Optimization","created":"2024-05-28","cited_by_count":118,"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":"2405.09719","title":"Spectral Editing of Activations for Large Language Model Alignment","created":"2024-05-15","cited_by_count":56,"area":"alignment"},{"id":"2408.11491","title":"SCANS: Mitigating the Exaggerated Safety for LLMs via Safety-Conscious Activation Steering","created":"2024-08-21","cited_by_count":50,"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":"2504.09466","title":"AdaSteer: Your Aligned LLM is Inherently an Adaptive Jailbreak Defender","created":"2025-04-13","cited_by_count":38,"area":"alignment"},{"id":"2408.11779","title":"Personality Alignment of Large Language Models","created":"2024-08-21","cited_by_count":36,"area":"alignment"},{"id":"2409.14586","title":"Backtracking Improves Generation Safety","created":"2024-09-22","cited_by_count":32,"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":"2412.00967","title":"Linear Probe Penalties Reduce LLM Sycophancy","created":"2024-12-01","cited_by_count":28,"area":"alignment"},{"id":"2407.08770","title":"Model Surgery: Modulating LLM's Behavior Via Simple Parameter Editing","created":"2024-07-11","cited_by_count":23,"area":"alignment"},{"id":"2602.17560","title":"ODESteer: A Unified ODE-Based Steering Framework for LLM Alignment","created":"2026-02-19","cited_by_count":19,"area":"alignment"},{"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":"2507.11316","title":"Internal Value Alignment in Large Language Models through Controlled Value Vector Activation","created":"2025-07-15","cited_by_count":16,"area":"alignment"},{"id":"2501.18532","title":"Differentially Private Steering for Large Language Model Alignment","created":"2025-01-30","cited_by_count":15,"area":"alignment"},{"id":"2506.00253","title":"Aligned but Blind: Alignment Increases Implicit Bias by Reducing Awareness of Race","created":"2025-05-30","cited_by_count":15,"area":"alignment"},{"id":"2510.04309","title":"Activation Steering with a Feedback Controller","created":"2025-10-05","cited_by_count":15,"area":"alignment"},{"id":"2505.20925","title":"Multi-objective Large Language Model Alignment with Hierarchical Experts","created":"2025-05-27","cited_by_count":13,"area":"alignment"},{"id":"2507.20333","title":"The Blessing and Curse of Dimensionality in Safety Alignment","created":"2025-07-27","cited_by_count":10,"area":"alignment"},{"id":"2510.01243","title":"Detoxifying Large Language Models via Autoregressive Reward Guided Representation Editing","created":"2025-09-24","cited_by_count":10,"area":"alignment"},{"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":"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":"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":"2501.10639","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":"2503.02989","title":"Effectively Steer LLM To Follow Preference via Building Confident Directions","created":"2025-03-04","cited_by_count":7,"area":"alignment"},{"id":"2604.01476","title":"From Rebound to Remedy: Understanding and Mitigating Reward Hacking via Representation Engineering","created":"2026-04-01","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":"2504.20106","title":"Adaptive Helpfulness-Harmlessness Alignment with Preference Vectors","created":"2025-04-27","cited_by_count":6,"area":"alignment"},{"id":"2411.06424","title":"How Does DPO Reduce Toxicity? A Mechanistic Neuron-Level Analysis","created":"2024-11-10","cited_by_count":5,"area":"alignment"},{"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":"2509.12934","title":"The Anatomy of Alignment: Decomposing Preference Optimization by Steering Sparse Features","created":"2025-09-16","cited_by_count":5,"area":"alignment"},{"id":"2512.23260","title":"Interpretable Safety Alignment via SAE-Constructed Low-Rank Subspace Adaptation","created":"2025-12-29","cited_by_count":5,"area":"alignment"},{"id":"2406.03642","title":"Is Free Self-Alignment Possible?","created":"2024-06-05","cited_by_count":4,"area":"alignment"},{"id":"2506.16078","title":"Probing the Robustness of Large Language Models Safety to Latent Perturbations","created":"2025-06-19","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":"2602.07639","title":"Letting Tutor Personas Speak Up for LLMs: Learning Steering Vectors from Dialogue via Preference Optimization","created":"2026-02-07","cited_by_count":4,"area":"alignment"},{"id":"2604.12359","title":"Compiling Activation Steering into Weights via Null-Space Constraints for Stealthy Backdoors","created":"2026-04-14","cited_by_count":4,"area":"alignment"},{"id":"2604.12479","title":"Meet Dynamic Individual Preferences: Resolving Conflicting Human Value with Paired Fine-Tuning","created":"2026-04-14","cited_by_count":4,"area":"alignment"},{"id":"2505.01162","title":"On the Limitations of Steering in Language Model Alignment","created":"2025-05-02","cited_by_count":3,"area":"alignment"},{"id":"2509.03647","title":"Breaking the Mirror: Activation-Based Mitigation of Self-Preference in LLM Evaluators","created":"2025-09-03","cited_by_count":3,"area":"alignment"},{"id":"2509.23982","title":"Toward Preference-aligned Large Language Models via Residual-based Model Steering","created":"2025-09-28","cited_by_count":3,"area":"alignment"},{"id":"2602.07356","title":"Controllable Value Alignment in Large Language Models through Neuron-Level Editing","created":"2026-02-07","cited_by_count":3,"area":"alignment"},{"id":"2603.03291","title":"One Bias After Another: Mechanistic Reward Shaping and Persistent Biases in Language Reward Models","created":"2026-02-06","cited_by_count":3,"area":"alignment"},{"id":"2604.03867","title":"Where to Steer: Input-Dependent Layer Selection for Steering Improves LLM Alignment","created":"2026-04-04","cited_by_count":3,"area":"alignment"},{"id":"2601.08441","title":"YaPO: Learnable Sparse Activation Steering Vectors for Domain Adaptation","created":"2026-01-13","cited_by_count":2,"area":"alignment"},{"id":"2603.03081","title":"TAO-Attack: Toward Advanced Optimization-Based Jailbreak Attacks for Large Language Models","created":"2026-03-03","cited_by_count":2,"area":"alignment"},{"id":"2603.05773","title":"Knowing without Acting: The Disentangled Geometry of Safety Mechanisms in Large Language Models","created":"2026-03-06","cited_by_count":2,"area":"alignment"},{"id":"2603.27518","title":"Over-Refusal and Representation Subspaces: A Mechanistic Analysis of Task-Conditioned Refusal in Aligned LLMs","created":"2026-03-29","cited_by_count":2,"area":"alignment"},{"id":"2609.19101","title":"Monitoring and Discovering Reward Hacking with Internal Representations during LLM Evaluations","created":"2026-09-16","cited_by_count":2,"area":"alignment"},{"id":"2509.16060","title":"SABER: Uncovering Vulnerabilities in Safety Alignment via Cross-Layer Residual Connection","created":"2025-09-19","cited_by_count":1,"area":"alignment"},{"id":"2601.12758","title":"VISPA: Pluralistic Alignment via Automatic Value Selection and Activation","created":"2026-01-19","cited_by_count":1,"area":"alignment"},{"id":"2602.19101","title":"Value Entanglement: Conflation Between Different Kinds of Good In (Some) Large Language Models","created":"2026-02-22","cited_by_count":1,"area":"alignment"},{"id":"2603.04355","title":"Efficient Refusal Ablation in LLM through Optimal Transport","created":"2026-03-04","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":"2603.21461","title":"DSPA: Dynamic SAE Steering for Data-Efficient Preference Alignment","created":"2026-03-23","cited_by_count":1,"area":"alignment"},{"id":"2604.27401","title":"Perturbation Probing: A Two-Pass-per-Prompt Diagnostic for FFN Behavioral Circuits in Aligned LLMs","created":"2026-04-30","cited_by_count":1,"area":"alignment"},{"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":"2605.17413","title":"Ablating Safety: Mechanisms for Removing Alignment in Language Models for Security Applications","created":"2026-05-17","cited_by_count":1,"area":"alignment"},{"id":"2605.21706","title":"Latent-space Attacks for Refusal Evasion in Language Models","created":"2026-05-20","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":"2606.09711","title":"Proxy Reward Internalization and Mechanistic Exploitation: A Learned Precursor to Reward Hacking and Its Generalization","created":"2026-06-08","cited_by_count":1,"area":"alignment"},{"id":"2607.01802","title":"On the Limits of Steering Vectors for Preference-Aligned Generation","created":"2026-07-02","cited_by_count":1,"area":"alignment"},{"id":"2607.02714","title":"Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale","created":"2026-07-02","cited_by_count":1,"area":"alignment"},{"id":"2607.12985","title":"Resist and Update: Counterfactual Report Coordinates for Incentive-Compatible LLMs","created":"2026-07-14","cited_by_count":1,"area":"alignment"},{"id":"2607.19243","title":"Inference-Time Steering for Cross-Lingual Factual Consistency in LLMs","created":"2026-07-21","cited_by_count":1,"area":"alignment"},{"id":"2609.05800","title":"Spillover-Aware Multi-Value Steering for Pluralistic LLM Alignment","created":"2026-09-05","cited_by_count":1,"area":"alignment"},{"id":"2508.06592","title":"Towards Integrated Alignment","created":"2025-08-08","cited_by_count":0,"area":"alignment"},{"id":"2601.10307","title":"The Straight and Narrow: Do LLMs Possess an Internal Moral Path?","created":"2026-01-15","cited_by_count":0,"area":"alignment"},{"id":"2603.03000","title":"Why Does RLAIF Work At All?","created":"2026-03-03","cited_by_count":0,"area":"alignment"},{"id":"2603.05839","title":"Evaluating LLM Alignment With Human Trust Models","created":"2026-03-06","cited_by_count":0,"area":"alignment"},{"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":"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":"2605.11712","title":"Toward Stable Value Alignment: Introducing Independent Modules for Consistent Value Guidance","created":"2026-05-12","cited_by_count":0,"area":"alignment"},{"id":"2605.23244","title":"Convex Optimization for Alignment and Preference Learning on a Single GPU","created":"2026-05-22","cited_by_count":0,"area":"alignment"},{"id":"2605.28722","title":"Multi-Adapter Representation Interventions via Energy Calibration","created":"2026-05-27","cited_by_count":0,"area":"alignment"},{"id":"2605.30526","title":"Measuring, Localizing, and Ablating Alignment Signatures in LLMs","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":"2606.19831","title":"Leverage Is Not Reach: A Control-Window Law for Single-Neuron Steering in Language Models","created":"2026-06-18","cited_by_count":0,"area":"alignment"},{"id":"2606.21645","title":"Behavioral and Representational Evidence of Binomial Ordering Preferences in Large Language Models","created":"2026-06-19","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":"2607.19371","title":"Mitigating Scaffolding Collapse in Socratic Tutors via Representation Alignment","created":"2026-06-15","cited_by_count":0,"area":"alignment"},{"id":"2608.03892","title":"Intertemporal Preference Steering in Qwen3 via Contrastive Activation Addition","created":"2026-08-04","cited_by_count":0,"area":"alignment"},{"id":"2608.07509","title":"PIVOT: Preference-based Intervention Vectors for Pedagogical Tutor Steering","created":"2026-06-29","cited_by_count":0,"area":"alignment"},{"id":"2608.26222","title":"NeuronFuzz: Safety Neuron Guided Fuzzing for LLM Safety Evaluation","created":"2026-08-26","cited_by_count":0,"area":"alignment"},{"id":"2608.30197","title":"ALTSTEER: Selective Safety Steering for Moving Beyond Hard Refusals to Constructive Alternatives","created":"2026-08-31","cited_by_count":0,"area":"alignment"},{"id":"2609.06289","title":"Steering Geometry: Validating Human Value Geometry in LLM Steering Space","created":"2026-09-05","cited_by_count":0,"area":"alignment"},{"id":"2609.06934","title":"The Geometry of Refusal: Why Post-Hoc Safety Is Fragile and Pretraining-Time Safety Persists","created":"2026-09-07","cited_by_count":0,"area":"alignment"},{"id":"2609.09793","title":"How Fragile Is Safety Alignment at Frontier Scale? A Single-Direction Attack on a 320B MoE","created":"2026-09-09","cited_by_count":0,"area":"alignment"},{"id":"2609.22359","title":"Resist, Update, Reject: Preference Optimization Installs a Prior-Dependent Reliability Switch","created":"2026-07-26","cited_by_count":0,"area":"alignment"},{"id":"2609.33298","title":"Direct Hidden-State Alignment: Mapping and Controlling Preference Expression in LLMs","created":"2026-09-27","cited_by_count":0,"area":"alignment"},{"id":"2609.34771","title":"When Do Model Internals Help? Exploring the Role of Representation Engineering in LLM Safety","created":"2026-09-28","cited_by_count":0,"area":"alignment"},{"id":"2610.19365","title":"DANGERLENS: Probing Dangerous Representations Hidden by Safety Alignment","created":"2026-10-14","cited_by_count":0,"area":"alignment"}],"next_cursor":null}