{"total":39,"rows":[{"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":"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":"2410.08847","title":"Unintentional Unalignment: Likelihood Displacement in Direct Preference Optimization","created":"2024-10-11","cited_by_count":67,"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":"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":"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":"2501.19358","title":"The Energy Loss Phenomenon in RLHF: A New Perspective on Mitigating Reward Hacking","created":"2025-01-31","cited_by_count":18,"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":"2510.04309","title":"Activation Steering with a Feedback Controller","created":"2025-10-05","cited_by_count":15,"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":"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":"2408.03459","title":"Can DPO Learn Diverse Human Values? A Theoretical Scaling Law","created":"2024-08-06","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":"2507.07981","title":"Why is Your Language Model a Poor Implicit Reward Model?","created":"2025-07-10","cited_by_count":6,"area":"alignment"},{"id":"2405.16964","title":"Exploring the LLM Journey from Cognition to Expression with Linear Representations","created":"2024-05-27","cited_by_count":5,"area":"alignment"},{"id":"2406.08124","title":"Legend: Leveraging Representation Engineering to Annotate Safety Margin for Preference Datasets","created":"2024-06-12","cited_by_count":5,"area":"alignment"},{"id":"2602.10623","title":"Mitigating Reward Hacking in RLHF via Bayesian Non-negative Reward Modeling","created":"2026-02-11","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":"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":"2501.06164","title":"Model Alignment Search","created":"2025-01-10","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":"2601.19231","title":"LLMs Can Unlearn Refusal with Only 1,000 Benign Samples","created":"2026-01-27","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":"2507.00239","title":"Linearly Decoding Refused Knowledge in Aligned Language Models","created":"2025-06-30","cited_by_count":2,"area":"alignment"},{"id":"2603.03326","title":"Controllable and explainable personality sliders for LLMs at inference time","created":"2026-02-10","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":"2606.21906","title":"Deeper is Not Always Better: Mitigating the Alignment Tax via Confident Layer Decoding","created":"2026-06-20","cited_by_count":2,"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":"2605.21706","title":"Latent-space Attacks for Refusal Evasion in Language Models","created":"2026-05-20","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":"2605.05995","title":"Safety Anchor: Defending Harmful Fine-tuning via Geometric Bottlenecks","created":"2026-05-07","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":"2605.28722","title":"Multi-Adapter Representation Interventions via Energy Calibration","created":"2026-05-27","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":"2608.26222","title":"NeuronFuzz: Safety Neuron Guided Fuzzing for LLM Safety Evaluation","created":"2026-08-26","cited_by_count":0,"area":"alignment"},{"id":"2609.32717","title":"LLM Alignment--Utility Asymmetry under Semantic-Preserving Transformations","created":"2026-09-26","cited_by_count":0,"area":"alignment"}],"next_cursor":null}