{"total":71,"rows":[{"id":"2410.14872","title":"How to Evaluate Reward Models for RLHF","created":"2024-10-18","cited_by_count":88,"area":"alignment"},{"id":"2506.11902","title":"TreeRL: LLM Reinforcement Learning with On-Policy Tree Search","created":"2025-06-13","cited_by_count":80,"area":"alignment"},{"id":"2412.14516","title":"Cal-DPO: Calibrated Direct Preference Optimization for Language Model Alignment","created":"2024-12-19","cited_by_count":51,"area":"alignment"},{"id":"2508.03686","title":"CompassVerifier: A Unified and Robust Verifier for LLMs Evaluation and Outcome Reward","created":"2025-08-05","cited_by_count":40,"area":"alignment"},{"id":"2503.11701","title":"A Survey of Direct Preference Optimization","created":"2025-03-12","cited_by_count":36,"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":"2506.19248","title":"Inference-Time Reward Hacking in Large Language Models","created":"2025-06-24","cited_by_count":34,"area":"alignment"},{"id":"2603.12273","title":"Aligning Language Models from User Interactions","created":"2026-02-18","cited_by_count":28,"area":"alignment"},{"id":"2502.16852","title":"Improving LLM General Preference Alignment via Optimistic Online Mirror Descent","created":"2025-02-24","cited_by_count":24,"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":"2501.12273","title":"Condor: Enhance LLM Alignment with Knowledge-Driven Data Synthesis and Refinement","created":"2025-01-21","cited_by_count":18,"area":"alignment"},{"id":"2507.05197","title":"Pre-Trained Policy Discriminators are General Reward Models","created":"2025-07-07","cited_by_count":15,"area":"alignment"},{"id":"2409.02231","title":"SmileyLlama: Modifying Large Language Models for Directed Chemical Space Exploration","created":"2024-09-03","cited_by_count":14,"area":"alignment"},{"id":"2510.01555","title":"Rethinking KL Regularization in RLHF: From Value Estimation to Gradient Optimization","created":"2025-10-02","cited_by_count":14,"area":"alignment"},{"id":"2412.09544","title":"Sail into the Headwind: Alignment via Robust Rewards and Dynamic Labels against Reward Hacking","created":"2024-12-12","cited_by_count":13,"area":"alignment"},{"id":"2503.05079","title":"On a Connection Between Imitation Learning and RLHF","created":"2025-03-07","cited_by_count":13,"area":"alignment"},{"id":"2505.10527","title":"WorldPM: Scaling Human Preference Modeling","created":"2025-05-15","cited_by_count":13,"area":"alignment"},{"id":"2510.17426","title":"Navigating the Alignment-Calibration Trade-off: A Pareto-Superior Frontier via Model Merging","created":"2025-10-20","cited_by_count":10,"area":"alignment"},{"id":"2604.02795","title":"Rubrics to Tokens: Bridging Response-level Rubrics and Token-level Rewards in Instruction Following Tasks","created":"2026-04-03","cited_by_count":10,"area":"alignment"},{"id":"2506.04463","title":"Aligning Large Language Models with Implicit Preferences from User-Generated Content","created":"2025-06-04","cited_by_count":9,"area":"alignment"},{"id":"2512.23457","title":"Replay Failures as Successes: Sample-Efficient Reinforcement Learning for Instruction Following","created":"2025-12-29","cited_by_count":8,"area":"alignment"},{"id":"2407.03856","title":"Q-Adapter: Customizing Pre-trained LLMs to New Preferences with Forgetting Mitigation","created":"2024-07-04","cited_by_count":7,"area":"alignment"},{"id":"2408.15339","title":"UNA: A Unified Supervised Framework for Efficient LLM Alignment Across Feedback Types","created":"2024-08-27","cited_by_count":6,"area":"alignment"},{"id":"2503.04222","title":"FuseChat-3.0: Preference Optimization Meets Heterogeneous Model Fusion","created":"2025-03-06","cited_by_count":6,"area":"alignment"},{"id":"2502.03397","title":"SPRI: Aligning Large Language Models with Context-Situated Principles","created":"2025-02-05","cited_by_count":5,"area":"alignment"},{"id":"2509.22099","title":"S2J: Bridging the Gap Between Solving and Judging Ability in Generative Reward Models","created":"2025-09-26","cited_by_count":5,"area":"alignment"},{"id":"2604.04215","title":"DARE: Diffusion Large Language Models Alignment and Reinforcement Executor","created":"2026-04-05","cited_by_count":5,"area":"alignment"},{"id":"2502.04675","title":"Scalable Oversight for Superhuman AI via Recursive Self-Critiquing","created":"2025-02-07","cited_by_count":4,"area":"alignment"},{"id":"2505.19327","title":"Paying Alignment Tax with Contrastive Learning","created":"2025-05-25","cited_by_count":3,"area":"alignment"},{"id":"2506.00835","title":"SynPO: Synergizing Descriptiveness and Preference Optimization for Video Detailed Captioning","created":"2025-06-01","cited_by_count":3,"area":"alignment"},{"id":"2506.07402","title":"Beyond Jailbreaks: Revealing Stealthier and Broader LLM Security Risks Stemming from Alignment Failures","created":"2025-06-09","cited_by_count":3,"area":"alignment"},{"id":"2510.20413","title":"Why DPO is a Misspecified Estimator and How to Fix It","created":"2025-10-23","cited_by_count":3,"area":"alignment"},{"id":"2602.08829","title":"WildReward: Learning Reward Models from In-the-Wild Human Interactions","created":"2026-02-09","cited_by_count":3,"area":"alignment"},{"id":"2604.17501","title":"CoAct: Co-Active LLM Preference Learning with Human-AI Synergy","created":"2026-04-19","cited_by_count":3,"area":"alignment"},{"id":"2606.11512","title":"SAGE: Answer-Conditioned Uncertainty Targets for Verbal Uncertainty Alignment","created":"2026-06-09","cited_by_count":3,"area":"alignment"},{"id":"2608.06377","title":"Learning When to Trust via Selective Context Preference Optimization","created":"2026-08-06","cited_by_count":3,"area":"alignment"},{"id":"2504.08590","title":"Playpen: An Environment for Exploring Learning Through Conversational Interaction","created":"2025-04-11","cited_by_count":2,"area":"alignment"},{"id":"2507.17107","title":"Reinforcement Learning Fine-Tunes a Sparse Subnetwork in Large Language Models","created":"2025-07-23","cited_by_count":2,"area":"alignment"},{"id":"2511.09385","title":"AMaPO: Adaptive Margin-attached Preference Optimization for Language Model Alignment","created":"2025-11-12","cited_by_count":2,"area":"alignment"},{"id":"2604.18239","title":"Towards Disentangled Preference Optimization Dynamics: Suppress the Loser, Preserve the Winner","created":"2026-04-20","cited_by_count":2,"area":"alignment"},{"id":"2608.19982","title":"COPA: Continual Preference Optimization for Adaptive Prompt Injection Defense","created":"2026-08-20","cited_by_count":2,"area":"alignment"},{"id":"2504.06562","title":"FuseRL: Dense Preference Optimization for Heterogeneous Model Fusion","created":"2025-04-09","cited_by_count":1,"area":"alignment"},{"id":"2509.25148","title":"AAPA: Adversarially Anchored Preference Alignment for Post-Training of Large Language Models","created":"2025-09-29","cited_by_count":1,"area":"alignment"},{"id":"2510.20187","title":"Every Question Has Its Own Value: Reinforcement Learning with Explicit Human Values","created":"2025-10-23","cited_by_count":1,"area":"alignment"},{"id":"2511.10985","title":"When Data is the Algorithm: A Systematic Study and Curation of Preference Optimization Datasets","created":"2025-11-14","cited_by_count":1,"area":"alignment"},{"id":"2602.03073","title":"TMS: Trajectory-Mixed Supervision for Reward-Free, On-Policy SFT","created":"2026-02-03","cited_by_count":1,"area":"alignment"},{"id":"2603.07211","title":"CompassDPO: Dynamics-Controlled Direct Preference Optimization for Robust Safety Alignment","created":"2026-03-07","cited_by_count":1,"area":"alignment"},{"id":"2604.08723","title":"Decomposing the Delta: What Do Models Actually Learn from Preference Pairs?","created":"2026-04-09","cited_by_count":1,"area":"alignment"},{"id":"2605.29582","title":"PEARL: Training Socratic Tutors with Pedagogically Aligned Reinforcement Learning","created":"2026-05-28","cited_by_count":1,"area":"alignment"},{"id":"2606.25442","title":"PolicyAlign: Direct Policy-Based Safety Alignment for Large Language Models","created":"2026-06-24","cited_by_count":1,"area":"alignment"},{"id":"2505.10775","title":"A Systematic Analysis of Base Model Choice for Reward Modeling","created":"2025-05-16","cited_by_count":0,"area":"alignment"},{"id":"2506.02018","title":"Enhancing Paraphrase Type Generation: The Impact of DPO and RLHF Evaluated with Human-Ranked Data","created":"2025-05-28","cited_by_count":0,"area":"alignment"},{"id":"2601.05570","title":"Crisis-Bench: Benchmarking Strategic Ambiguity and Reputation Management in Large Language Models","created":"2026-01-09","cited_by_count":0,"area":"alignment"},{"id":"2602.02898","title":"Aligning Language Model Benchmarks with Pairwise Preferences","created":"2026-02-02","cited_by_count":0,"area":"alignment"},{"id":"2603.20217","title":"Expected Reward Prediction, with Applications to Model Routing","created":"2026-03-03","cited_by_count":0,"area":"alignment"},{"id":"2604.15602","title":"GroupDPO: Memory-Efficient Group-Wise Direct Preference Optimization","created":"2026-04-17","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":"2605.20654","title":"REFLECTOR: Internalizing Step-wise Reflection against Indirect Jailbreak","created":"2026-05-20","cited_by_count":0,"area":"alignment"},{"id":"2605.26293","title":"CroCo: Cross-Lingual Contrastive Preference Tuning on Self-Generations","created":"2026-05-25","cited_by_count":0,"area":"alignment"},{"id":"2606.08032","title":"Variational Proximal Policy Optimization","created":"2026-06-06","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":"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":"2609.03960","title":"FiMI Banking: A Sovereign Model for Indian Retail Banking","created":"2026-09-03","cited_by_count":0,"area":"alignment"},{"id":"2609.04022","title":"Representational alignment yields generalizable safety in language models","created":"2026-09-03","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":"2609.32189","title":"ScopeIF: Improving Scope-Aware Precise Instruction-Following in Large Language Models via Graded Reward Modeling","created":"2026-09-26","cited_by_count":0,"area":"alignment"},{"id":"2610.04328","title":"Bidirectional Preference Synthesis: Learning Prompt-Conditioned Preferences from Boundary Failures","created":"2026-10-03","cited_by_count":0,"area":"alignment"},{"id":"2610.07935","title":"SIGMA: Self-Improving Alignment Generalization from a Model Spec","created":"2026-10-06","cited_by_count":0,"area":"alignment"},{"id":"2610.12233","title":"ReSI: Recursive Safety Improvement toward Resistant and Resilient AI","created":"2026-10-08","cited_by_count":0,"area":"alignment"},{"id":"2610.14578","title":"Uniform Information Density-Based Preference Optimization","created":"2026-10-10","cited_by_count":0,"area":"alignment"},{"id":"2610.20350","title":"Lifted State Policy Optimization: Internalizing Self-Correction through Transition Geometry","created":"2026-10-14","cited_by_count":0,"area":"alignment"}],"next_cursor":null}