{"total":95,"rows":[{"id":"2410.08847","title":"Unintentional Unalignment: Likelihood Displacement in Direct Preference Optimization","created":"2024-10-11","cited_by_count":67,"area":"alignment"},{"id":"2506.09026","title":"e3: Learning to Explore Enables Extrapolation of Test-Time Compute for LLMs","created":"2025-06-10","cited_by_count":59,"area":"lr"},{"id":"2505.12929","title":"Do Not Let Low-Probability Tokens Over-Dominate in RL for LLMs","created":"2025-05-19","cited_by_count":44,"area":"lr"},{"id":"2602.04879","title":"Rethinking the Trust Region in LLM Reinforcement Learning","created":"2026-02-04","cited_by_count":37,"area":"lr"},{"id":"2509.23024","title":"Tracing the Representation Geometry of Language Models from Pretraining to Post-training","created":"2025-09-27","cited_by_count":35,"area":null},{"id":"2505.18830","title":"On the Effect of Negative Gradient in Group Relative Deep Reinforcement Optimization","created":"2025-05-24","cited_by_count":31,"area":"lr"},{"id":"2604.21691","title":"There Will Be a Scientific Theory of Deep Learning","created":"2026-04-23","cited_by_count":29,"area":null},{"id":"2410.15595","title":"A Comprehensive Survey of Direct Preference Optimization: Datasets, Theories, Variants, and Applications","created":"2024-10-21","cited_by_count":27,"area":"lr"},{"id":"2507.23779","title":"Phi-Ground Tech Report: Advancing Perception in GUI Grounding","created":"2025-07-31","cited_by_count":25,"area":"lr"},{"id":"2505.04560","title":"ABKD: Pursuing a Proper Allocation of the Probability Mass in Knowledge Distillation via $\\alpha$-$\\beta$-Divergence","created":"2025-05-07","cited_by_count":23,"area":"distill"},{"id":"2510.08233","title":"Enhancing Reasoning for Diffusion LLMs via Distribution Matching Policy Optimization","created":"2025-10-09","cited_by_count":20,"area":"lr"},{"id":"2510.02230","title":"The Reasoning Boundary Paradox: How Reinforcement Learning Constrains Language Models","created":"2025-10-02","cited_by_count":18,"area":"lr"},{"id":"2507.00018","title":"Implicit Reward as the Bridge: A Unified View of SFT and DPO Connections","created":"2025-06-15","cited_by_count":17,"area":"alignment"},{"id":"2510.19422","title":"LLM Unlearning with LLM Beliefs","created":"2025-10-22","cited_by_count":17,"area":"safety"},{"id":"2507.21183","title":"MaPPO: Maximum a Posteriori Preference Optimization with Prior Knowledge","created":"2025-07-27","cited_by_count":15,"area":"alignment"},{"id":"2512.04220","title":"On Group Relative Policy Optimization Collapse in Agent Search: The Lazy Likelihood-Displacement","created":"2025-12-03","cited_by_count":15,"area":"lr"},{"id":"2506.23508","title":"Why Reinforcement Fine-Tuning Enables MLLMs Preserve Prior Knowledge Better: A Data Perspective","created":"2025-06-30","cited_by_count":14,"area":"lr"},{"id":"2411.14491","title":"A Survey on Human-Centric LLMs","created":"2024-11-20","cited_by_count":13,"area":null},{"id":"2505.05465","title":"ComPO: Preference Alignment via Comparison Oracles","created":"2025-05-08","cited_by_count":13,"area":"alignment"},{"id":"2508.09494","title":"Learning Facts at Scale with Active Reading","created":"2025-08-13","cited_by_count":13,"area":null},{"id":"2510.04028","title":"The Debate on RLVR Reasoning Capability Boundary: Shrinkage, Expansion, or Both? A Two-Stage Dynamic View","created":"2025-10-05","cited_by_count":13,"area":"lr"},{"id":"2603.22117","title":"On the Direction of RLVR Updates for LLM Reasoning: Identification and Exploitation","created":"2026-03-23","cited_by_count":13,"area":"lr"},{"id":"2505.12504","title":"CPGD: Toward Stable Rule-based Reinforcement Learning for Language Models","created":"2025-05-18","cited_by_count":11,"area":"lr"},{"id":"2510.03669","title":"Token Hidden Reward: Steering Exploration-Exploitation in Group Relative Deep Reinforcement Learning","created":"2025-10-04","cited_by_count":10,"area":"lr"},{"id":"2510.26707","title":"Value Drifts: Tracing Value Alignment During LLM Post-Training","created":"2025-10-30","cited_by_count":10,"area":"alignment"},{"id":"2509.21044","title":"Reinforcement Learning Fine-Tuning Enhances Activation Intensity and Diversity in the Internal Circuitry of LLMs","created":"2025-09-25","cited_by_count":9,"area":null},{"id":"2505.18917","title":"Behavior Injection: Preparing Language Models for Reinforcement Learning","created":"2025-05-25","cited_by_count":8,"area":"lr"},{"id":"2509.26313","title":"One-Token Rollout: Guiding Supervised Fine-Tuning of LLMs with Policy Gradient","created":"2025-09-30","cited_by_count":8,"area":null},{"id":"2602.03392","title":"On the Entropy Dynamics in Reinforcement Fine-Tuning of Large Language Models","created":"2026-02-03","cited_by_count":8,"area":null},{"id":"2603.06495","title":"COLD-Steer: Steering Large Language Models via In-Context One-step Learning Dynamics","created":"2026-03-06","cited_by_count":8,"area":"mechinterp"},{"id":"2505.12306","title":"Bidirectional LMs are Better Knowledge Memorizers? A Benchmark for Real-world Knowledge Injection","created":"2025-05-18","cited_by_count":7,"area":null},{"id":"2510.03817","title":"TROLL: Trust Regions improve Reinforcement Learning for Large Language Models","created":"2025-10-04","cited_by_count":7,"area":null},{"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":"2509.23383","title":"Train Once, Answer All: Many Pretraining Experiments for the Cost of One","created":"2025-09-27","cited_by_count":6,"area":null},{"id":"2510.16629","title":"On the Impossibility of Retrain Equivalence in Machine Unlearning","created":"2025-10-18","cited_by_count":6,"area":"safety"},{"id":"2603.00498","title":"Antibody: Strengthening Defense Against Harmful Fine-Tuning for Large Language Models via Attenuating Harmful Gradient Influence","created":"2026-02-28","cited_by_count":6,"area":"safety"},{"id":"2505.23316","title":"Proximalized Preference Optimization for Diverse Feedback Types: A Decomposed Perspective on DPO","created":"2025-05-29","cited_by_count":5,"area":"alignment"},{"id":"2509.20868","title":"StyleBench: Evaluating thinking styles in Large Language Models","created":"2025-09-25","cited_by_count":5,"area":"lr"},{"id":"2602.05717","title":"Anchored Policy Optimization: Mitigating Exploration Collapse Via Support-Constrained Rectification","created":"2026-02-05","cited_by_count":5,"area":null},{"id":"2604.20051","title":"Bootstrapping Post-training Signals for Open-ended Tasks via Rubric-based Self-play on Pre-training Text","created":"2026-04-21","cited_by_count":5,"area":"synth"},{"id":"2608.14728","title":"Tail-Aware Top-$k$ On-Policy Distillation","created":"2026-08-12","cited_by_count":5,"area":"distill"},{"id":"2502.17721","title":"Aligning Compound AI Systems via System-level DPO","created":"2025-02-24","cited_by_count":4,"area":"alignment"},{"id":"2505.17859","title":"Scalable Valuation of Human Feedback through Provably Robust Model Alignment","created":"2025-05-23","cited_by_count":4,"area":"alignment"},{"id":"2511.07368","title":"Distributional Biases in Post-Training: A Markovian Analysis of Reasoning Trajectories","created":"2025-11-10","cited_by_count":4,"area":"lr"},{"id":"2512.00706","title":"Optimizing LVLMs with On-Policy Data for Effective Hallucination Mitigation","created":"2025-11-30","cited_by_count":4,"area":"mllm"},{"id":"2604.16279","title":"Evaluating the Progression of Large Language Model Capabilities for Small-Molecule Drug Design","created":"2026-04-17","cited_by_count":4,"area":null},{"id":"2605.02944","title":"Exploring Pass-Rate Reward in Reinforcement Learning for Code Generation","created":"2026-05-01","cited_by_count":4,"area":null},{"id":"2606.19950","title":"Confidence Calibration for Multimodal LLMs: An Empirical Study through Medical VQA","created":"2026-06-18","cited_by_count":4,"area":"mllm"},{"id":"2502.06106","title":"Fine-Tuning is Subgraph Search: A New Lens on Learning Dynamics","created":"2025-02-10","cited_by_count":3,"area":"mechinterp"},{"id":"2505.17988","title":"Towards Revealing the Effectiveness of Small-Scale Fine-tuning in R1-style Reinforcement Learning","created":"2025-05-23","cited_by_count":3,"area":"lr"},{"id":"2508.20230","title":"Coresets from Trajectories: Selecting Data via Correlation of Loss Differences","created":"2025-08-27","cited_by_count":3,"area":null},{"id":"2509.23166","title":"Test-Time Policy Adaptation for Enhanced Multi-Turn Interactions with LLMs","created":"2025-09-27","cited_by_count":3,"area":null},{"id":"2603.01683","title":"Surgical Post-Training: Proximal On-Policy Distillation for Reasoning with Knowledge Retention","created":"2026-03-02","cited_by_count":3,"area":"lr"},{"id":"2603.03323","title":"Discern Truth from Falsehood: Reducing Over-Refusal via Contrastive Refinement","created":"2026-02-10","cited_by_count":3,"area":"safety"},{"id":"2605.08666","title":"The Cancellation Hypothesis in Critic-Free RL: From Outcome Rewards to Token Credits","created":"2026-05-09","cited_by_count":3,"area":null},{"id":"2609.20807","title":"Score Centering Stabilizes Off-policy Reinforcement Learning","created":"2026-09-17","cited_by_count":3,"area":"lr"},{"id":"2504.13125","title":"LLMs Meet Finance: Fine-Tuning Foundation Models for the Open FinLLM Leaderboard","created":"2025-04-17","cited_by_count":2,"area":null},{"id":"2506.00175","title":"Who Gets Credit or Blame? Attributing Accountability in Modern AI Systems","created":"2025-05-30","cited_by_count":2,"area":null},{"id":"2509.25771","title":"Free Lunch Alignment of Text-to-Image Diffusion Models without Preference Image Pairs","created":"2025-09-30","cited_by_count":2,"area":"diffusion"},{"id":"2511.21056","title":"Bilevel Data Curation for LLM Fine-tuning: Offline Selection and Online Self-Refining Generation","created":"2025-11-26","cited_by_count":2,"area":null},{"id":"2512.00778","title":"What Is Preference Optimization Doing, and Why?","created":"2025-11-30","cited_by_count":2,"area":"alignment"},{"id":"2601.04670","title":"Learning Dynamics in RL Post-Training for Language Models","created":"2026-01-08","cited_by_count":2,"area":"lr"},{"id":"2604.15705","title":"Towards Robust Endogenous Reasoning: Unifying Drift Adaptation in Non-Stationary Tuning","created":"2026-04-17","cited_by_count":2,"area":"lr"},{"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":"2605.06523","title":"On the Implicit Reward Overfitting and the Low-rank Dynamics in RLVR","created":"2026-05-07","cited_by_count":2,"area":"lr"},{"id":"2605.18309","title":"Alignment Dynamics in LLM Fine-Tuning","created":"2026-05-18","cited_by_count":2,"area":"alignment"},{"id":"2606.06667","title":"The Piggyback Hypothesis of Generalization: Explaining and Mitigating Emergent Misalignment","created":"2026-06-04","cited_by_count":2,"area":null},{"id":"2607.22039","title":"Enough is as good as a feast: A Comprehensive Analysis of How Reinforcement Learning Mitigates Task Conflicts in LLMs","created":"2026-07-24","cited_by_count":2,"area":null},{"id":"2608.22631","title":"Learning Generalizable Behaviors for Terminal Agents","created":"2026-08-23","cited_by_count":2,"area":"agents"},{"id":"2505.16875","title":"T2I-ConBench: Text-to-Image Benchmark for Continual Post-training","created":"2025-05-22","cited_by_count":1,"area":"diffusion"},{"id":"2507.01752","title":"Tuning without Peeking: Provable Generalization Bounds and Robust LLM Post-Training","created":"2025-07-02","cited_by_count":1,"area":"safety"},{"id":"2509.00183","title":"FNODE: Flow-Matching for data-driven simulation of constrained multibody systems","created":"2025-08-29","cited_by_count":1,"area":"diffusion"},{"id":"2509.06239","title":"Proof2Silicon: Prompt Repair for Verified Code and Hardware Generation via Reinforcement Learning","created":"2025-09-07","cited_by_count":1,"area":"lr"},{"id":"2603.18258","title":"Sharpness-Aware Minimization in Logit Space Efficiently Enhances Direct Preference Optimization","created":"2026-03-18","cited_by_count":1,"area":"alignment"},{"id":"2604.04894","title":"Asymmetric Advantage Modulation Calibrates Entropy Dynamics in RLVR","created":"2026-04-06","cited_by_count":1,"area":"lr"},{"id":"2604.26326","title":"Addressing Performance Saturation for LLM RL via Precise Entropy Curve Control","created":"2026-04-29","cited_by_count":1,"area":"lr"},{"id":"2605.02626","title":"Gradient-Gated DPO: Stabilizing Preference Optimization in Language Models","created":"2026-05-04","cited_by_count":1,"area":"alignment"},{"id":"2605.03373","title":"Learning Dynamics of Zeroth-Order Optimization: A Kernel Perspective","created":"2026-05-05","cited_by_count":1,"area":null},{"id":"2605.07353","title":"Confidence-Aware Alignment Makes Reasoning LLMs More Reliable","created":"2026-05-08","cited_by_count":1,"area":"lr"},{"id":"2605.27774","title":"Fine-Tuning Dynamics of In-Context Factual Recall in Transformers","created":"2026-05-26","cited_by_count":1,"area":null},{"id":"2606.01558","title":"Attention-guided Fine-tuning of Multimodal Large Language Models Improves Chain-of-Thought Reasoning","created":"2026-06-01","cited_by_count":1,"area":"lr"},{"id":"2505.21893","title":"SIPO: Stabilized and Improved Preference Optimization for Aligning Diffusion Models","created":"2025-05-28","cited_by_count":0,"area":"diffusion"},{"id":"2509.04442","title":"Delta Activations: A Representation for Finetuned Large Language Models","created":"2025-09-04","cited_by_count":0,"area":null},{"id":"2605.04572","title":"From Parameter Dynamics to Risk Scoring : Quantifying Sample-Level Safety Degradation in LLM Fine-tuning","created":"2026-05-06","cited_by_count":0,"area":"alignment"},{"id":"2606.26857","title":"LCAi: Life Cycle Assessment with big data fusion and retrieval-augmented generation-assisted interpretation","created":"2026-06-25","cited_by_count":0,"area":null},{"id":"2608.30699","title":"Learning Dynamics of Logits Debiasing for Long-Tailed Semi-Supervised Learning","created":"2026-08-31","cited_by_count":0,"area":null},{"id":"2609.03241","title":"FlowBalance: Verifier-Grounded Self-Improvement from On-Policy Reasoning Experience","created":"2026-09-03","cited_by_count":0,"area":"lr"},{"id":"2609.19144","title":"A Zeroth-Order Paradigm for LLM Preference Alignment","created":"2026-09-16","cited_by_count":0,"area":"alignment"},{"id":"2609.23260","title":"Why Ghost Outputs Teach: A Kernel-Based Understanding of Subliminal Learning","created":"2026-09-20","cited_by_count":0,"area":null},{"id":"2609.31214","title":"Which Influence Are We Estimating? The Role of Counterfactual Specifications in Data Attribution","created":"2026-09-25","cited_by_count":0,"area":null},{"id":"2609.33620","title":"Learning Dynamics of Continual Learning: A Unified View of Data Attribution, Forgetting, and Plasticity Loss","created":"2026-09-27","cited_by_count":0,"area":null},{"id":"2609.33781","title":"Surprising Success, Repeated Failure: Entropy-Guided Credit Assignment for Exploration in LLM Reasoning","created":"2026-09-27","cited_by_count":0,"area":"lr"},{"id":"2609.34849","title":"When Sparse Reward Meets Dense Distillation: Training Dynamics of On-Policy Distillation","created":"2026-09-28","cited_by_count":0,"area":"distill"},{"id":"2610.05373","title":"Towards Unbiased On-Policy Distillation for Block Diffusion Language Models","created":"2026-10-04","cited_by_count":0,"area":"diffusion"},{"id":"2610.09164","title":"Training Language Models To Be Coherent Decision-Makers","created":"2026-10-06","cited_by_count":0,"area":null}],"next_cursor":null}