{"total":225,"rows":[{"id":"2212.08073","title":"Constitutional AI: Harmlessness from AI Feedback","created":"2022-12-15","cited_by_count":3050,"area":"lr"},{"id":"2311.12022","title":"GPQA: A Graduate-Level Google-Proof Q&A Benchmark","created":"2023-11-20","cited_by_count":2879,"area":null},{"id":"2310.13548","title":"Towards Understanding Sycophancy in Language Models","created":"2023-10-20","cited_by_count":1040,"area":null},{"id":"2212.09251","title":"Discovering Language Model Behaviors with Model-Written Evaluations","created":"2022-12-19","cited_by_count":836,"area":"alignment"},{"id":"2307.15217","title":"Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback","created":"2023-07-27","cited_by_count":759,"area":"alignment"},{"id":"2307.13702","title":"Measuring Faithfulness in Chain-of-Thought Reasoning","created":"2023-07-17","cited_by_count":703,"area":"lr"},{"id":"2312.09390","title":"Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision","created":"2023-12-14","cited_by_count":467,"area":"alignment"},{"id":"2308.06463","title":"GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher","created":"2023-08-12","cited_by_count":423,"area":"alignment"},{"id":"2304.13712","title":"Harnessing the Power of LLMs in Practice: A Survey on ChatGPT and Beyond","created":"2023-04-26","cited_by_count":411,"area":null},{"id":"2310.10631","title":"Llemma: An Open Language Model For Mathematics","created":"2023-10-16","cited_by_count":388,"area":"lr"},{"id":"2503.11926","title":"Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation","created":"2025-03-14","cited_by_count":384,"area":"lr"},{"id":"2412.06559","title":"ProcessBench: Identifying Process Errors in Mathematical Reasoning","created":"2024-12-09","cited_by_count":292,"area":"lr"},{"id":"2402.06782","title":"Debating with More Persuasive LLMs Leads to More Truthful Answers","created":"2024-02-09","cited_by_count":266,"area":"alignment"},{"id":"2309.15025","title":"Large Language Model Alignment: A Survey","created":"2023-09-26","cited_by_count":243,"area":"alignment"},{"id":"2312.06942","title":"AI Control: Improving Safety Despite Intentional Subversion","created":"2023-12-12","cited_by_count":229,"area":"safety"},{"id":"2305.10142","title":"Improving Language Model Negotiation with Self-Play and In-Context Learning from AI Feedback","created":"2023-05-17","cited_by_count":207,"area":"synth"},{"id":"2407.19594","title":"Meta-Rewarding Language Models: Self-Improving Alignment with LLM-as-a-Meta-Judge","created":"2024-07-28","cited_by_count":204,"area":"synth"},{"id":"2307.10928","title":"FLASK: Fine-grained Language Model Evaluation based on Alignment Skill Sets","created":"2023-07-20","cited_by_count":201,"area":"alignment"},{"id":"2308.03958","title":"Simple synthetic data reduces sycophancy in large language models","created":"2023-08-07","cited_by_count":187,"area":"synth"},{"id":"2402.05070","title":"A Roadmap to Pluralistic Alignment","created":"2024-02-07","cited_by_count":184,"area":"alignment"},{"id":"2401.12187","title":"WARM: On the Benefits of Weight Averaged Reward Models","created":"2024-01-22","cited_by_count":147,"area":"alignment"},{"id":"2407.00215","title":"LLM Critics Help Catch LLM Bugs","created":"2024-06-28","cited_by_count":147,"area":"lr"},{"id":"2502.21321","title":"LLM Post-Training: A Deep Dive into Reasoning Large Language Models","created":"2025-02-28","cited_by_count":135,"area":"lr"},{"id":"2404.12253","title":"Toward Self-Improvement of LLMs via Imagination, Searching, and Criticizing","created":"2024-04-18","cited_by_count":125,"area":"lr"},{"id":"2402.07510","title":"Secret Collusion among AI Agents: Multi-Agent Deception via Steganography","created":"2024-02-12","cited_by_count":113,"area":"agents"},{"id":"2404.07503","title":"Best Practices and Lessons Learned on Synthetic Data","created":"2024-04-11","cited_by_count":106,"area":"synth"},{"id":"2407.04622","title":"On scalable oversight with weak LLMs judging strong LLMs","created":"2024-07-05","cited_by_count":99,"area":null},{"id":"2307.03718","title":"Frontier AI Regulation: Managing Emerging Risks to Public Safety","created":"2023-07-06","cited_by_count":98,"area":null},{"id":"2403.09472","title":"Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision","created":"2024-03-14","cited_by_count":96,"area":"alignment"},{"id":"2212.09746","title":"Evaluating Human-Language Model Interaction","created":"2022-12-19","cited_by_count":92,"area":null},{"id":"2303.05453","title":"Personalisation within bounds: A risk taxonomy and policy framework for the alignment of large language models with personalised feedback","created":"2023-03-09","cited_by_count":87,"area":"alignment"},{"id":"2407.13692","title":"Prover-Verifier Games improve legibility of LLM outputs","created":"2024-07-18","cited_by_count":85,"area":null},{"id":"2406.04268","title":"Open-Endedness is Essential for Artificial Superhuman Intelligence","created":"2024-06-06","cited_by_count":80,"area":null},{"id":"2303.16749","title":"Improving Code Generation by Training with Natural Language Feedback","created":"2023-03-28","cited_by_count":76,"area":null},{"id":"2312.01037","title":"Eliciting Latent Knowledge from Quirky Language Models","created":"2023-12-02","cited_by_count":62,"area":null},{"id":"2410.02052","title":"ExACT: Teaching AI Agents to Explore with Reflective-MCTS and Exploratory Learning","created":"2024-10-02","cited_by_count":62,"area":"lr"},{"id":"2407.21792","title":"Safetywashing: Do AI Safety Benchmarks Actually Measure Safety Progress?","created":"2024-07-31","cited_by_count":61,"area":null},{"id":"2408.16737","title":"Smaller, Weaker, Yet Better: Training LLM Reasoners via Compute-Optimal Sampling","created":"2024-08-29","cited_by_count":61,"area":"lr"},{"id":"2411.01855","title":"Can Language Models Learn to Skip Steps?","created":"2024-11-04","cited_by_count":59,"area":"lr"},{"id":"2310.05910","title":"SALMON: Self-Alignment with Instructable Reward Models","created":"2023-10-09","cited_by_count":55,"area":"alignment"},{"id":"2310.07629","title":"The Past, Present and Better Future of Feedback Learning in Large Language Models for Subjective Human Preferences and Values","created":"2023-10-11","cited_by_count":54,"area":"alignment"},{"id":"2502.14127","title":"Which of These Best Describes Multiple Choice Evaluation with LLMs? A) Forced B) Flawed C) Fixable D) All of the Above","created":"2025-02-19","cited_by_count":52,"area":null},{"id":"2305.07722","title":"In Search of Verifiability: Explanations Rarely Enable Complementary Performance in AI-Advised Decision Making","created":"2023-05-12","cited_by_count":50,"area":null},{"id":"2405.19107","title":"Offline Regularised Reinforcement Learning for Large Language Models Alignment","created":"2024-05-29","cited_by_count":48,"area":"alignment"},{"id":"2306.09983","title":"Evaluating Superhuman Models with Consistency Checks","created":"2023-06-16","cited_by_count":47,"area":null},{"id":"2310.13798","title":"Specific versus General Principles for Constitutional AI","created":"2023-10-20","cited_by_count":47,"area":"alignment"},{"id":"2311.10538","title":"Testing Language Model Agents Safely in the Wild","created":"2023-11-17","cited_by_count":46,"area":"agents"},{"id":"2504.10374","title":"Ctrl-Z: Controlling AI Agents via Resampling","created":"2025-04-14","cited_by_count":44,"area":null},{"id":"2311.08702","title":"Debate Helps Supervise Unreliable Experts","created":"2023-11-15","cited_by_count":43,"area":null},{"id":"2311.07954","title":"A Closer Look at the Self-Verification Abilities of Large Language Models in Logical Reasoning","created":"2023-11-14","cited_by_count":41,"area":"lr"},{"id":"2403.05518","title":"Bias-Augmented Consistency Training Reduces Biased Reasoning in Chain-of-Thought","created":"2024-03-08","cited_by_count":39,"area":"lr"},{"id":"2411.16345","title":"Preference Optimization for Reasoning with Pseudo Feedback","created":"2024-11-25","cited_by_count":39,"area":"lr"},{"id":"2507.16795","title":"Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning","created":"2025-07-22","cited_by_count":39,"area":null},{"id":"2504.03846","title":"Do LLM Evaluators Prefer Themselves for a Reason?","created":"2025-04-04","cited_by_count":38,"area":"lr"},{"id":"2407.21009","title":"AI-Assisted Generation of Difficult Math Questions","created":"2024-07-30","cited_by_count":37,"area":"lr"},{"id":"2408.11865","title":"How Susceptible are LLMs to Influence in Prompts?","created":"2024-08-17","cited_by_count":33,"area":null},{"id":"2604.13602","title":"Reward Hacking in the Era of Large Models: Mechanisms, Emergent Misalignment, Challenges","created":"2026-04-15","cited_by_count":33,"area":"alignment"},{"id":"2405.10313","title":"How Far Are We From AGI: Are LLMs All We Need?","created":"2024-05-16","cited_by_count":31,"area":null},{"id":"2508.19461","title":"Reliable Weak-to-Strong Monitoring of LLM Agents","created":"2025-08-26","cited_by_count":31,"area":"agents"},{"id":"2504.21848","title":"Characterizing AI Agents for Alignment and Governance","created":"2025-04-30","cited_by_count":30,"area":null},{"id":"2507.06187","title":"The Delta Learning Hypothesis: Preference Tuning on Weak Data can Yield Strong Gains","created":"2025-07-08","cited_by_count":28,"area":null},{"id":"2510.01367","title":"Is It Thinking or Cheating? Detecting Implicit Reward Hacking by Measuring Reasoning Effort","created":"2025-10-01","cited_by_count":27,"area":"lr"},{"id":"2310.13011","title":"Compositional preference models for aligning LMs","created":"2023-10-17","cited_by_count":26,"area":"alignment"},{"id":"2401.11458","title":"Linear Alignment: A Closed-form Solution for Aligning Human Preferences without Tuning and Feedback","created":"2024-01-21","cited_by_count":25,"area":"alignment"},{"id":"2406.11191","title":"A Survey on Human Preference Learning for Large Language Models","created":"2024-06-17","cited_by_count":24,"area":"alignment"},{"id":"2601.20245","title":"How AI Impacts Skill Formation","created":"2026-01-28","cited_by_count":24,"area":null},{"id":"2310.02932","title":"Assessing Large Language Models on Climate Information","created":"2023-10-04","cited_by_count":23,"area":null},{"id":"2504.18530","title":"Scaling Laws For Scalable Oversight","created":"2025-04-25","cited_by_count":23,"area":"safety"},{"id":"2311.08576","title":"Towards Evaluating AI Systems for Moral Status Using Self-Reports","created":"2023-11-14","cited_by_count":22,"area":null},{"id":"2410.09302","title":"Enhancing Multi-Step Reasoning Abilities of Language Models through Direct Q-Function Optimization","created":"2024-10-11","cited_by_count":22,"area":"lr"},{"id":"2403.04204","title":"On the Essence and Prospect: An Investigation of Alignment Approaches for Big Models","created":"2024-03-07","cited_by_count":21,"area":"alignment"},{"id":"2506.22777","title":"Teaching Models to Verbalize Reward Hacking in Chain-of-Thought Reasoning","created":"2025-06-28","cited_by_count":21,"area":"lr"},{"id":"2310.04815","title":"Critique Ability of Large Language Models","created":"2023-10-07","cited_by_count":20,"area":null},{"id":"2311.01007","title":"Effective Human-AI Teams via Learned Natural Language Rules and Onboarding","created":"2023-11-02","cited_by_count":20,"area":null},{"id":"2602.11865","title":"Intelligent AI Delegation","created":"2026-02-12","cited_by_count":18,"area":null},{"id":"2405.19524","title":"AI Risk Management Should Incorporate Both Safety and Security","created":"2024-05-29","cited_by_count":17,"area":null},{"id":"2505.00662","title":"DeepCritic: Deliberate Critique with Large Language Models","created":"2025-05-01","cited_by_count":17,"area":null},{"id":"2401.04620","title":"Agent Alignment in Evolving Social Norms","created":"2024-01-09","cited_by_count":15,"area":"agents"},{"id":"2405.16236","title":"A transfer learning framework for weak-to-strong generalization","created":"2024-05-25","cited_by_count":14,"area":null},{"id":"2310.17639","title":"In-Context Learning Dynamics with Random Binary Sequences","created":"2023-10-26","cited_by_count":13,"area":null},{"id":"2402.13764","title":"CriticEval: Evaluating Large Language Model as Critic","created":"2024-02-21","cited_by_count":13,"area":null},{"id":"2412.16468","title":"The Road to Artificial SuperIntelligence: A Comprehensive Survey of Superalignment","created":"2024-12-21","cited_by_count":13,"area":"alignment"},{"id":"2502.15861","title":"C3AI: Crafting and Evaluating Constitutions for Constitutional AI","created":"2025-02-21","cited_by_count":13,"area":"alignment"},{"id":"2410.13502","title":"MathGAP: Out-of-Distribution Evaluation on Problems with Arbitrarily Complex Proofs","created":"2024-10-17","cited_by_count":12,"area":null},{"id":"2501.05727","title":"Self-Evolving Critique Abilities in Large Language Models","created":"2025-01-10","cited_by_count":12,"area":"lr"},{"id":"2407.09221","title":"Evaluating AI Evaluation: Perils and Prospects","created":"2024-07-12","cited_by_count":11,"area":null},{"id":"2412.09572","title":"Rethinking LLM Uncertainty: A Multi-Agent Approach to Estimating Black-Box Model Uncertainty","created":"2024-12-12","cited_by_count":11,"area":"lr"},{"id":"2405.10295","title":"Societal Adaptation to Advanced AI","created":"2024-05-16","cited_by_count":10,"area":null},{"id":"2502.07577","title":"Automated Capability Discovery via Foundation Model Self-Exploration","created":"2025-02-11","cited_by_count":10,"area":null},{"id":"2505.01482","title":"Understanding LLM Scientific Reasoning through Promptings and Model's Explanation on the Answers","created":"2025-05-02","cited_by_count":10,"area":"lr"},{"id":"2506.05296","title":"Control Tax: The Price of Keeping AI in Check","created":"2025-06-05","cited_by_count":10,"area":"agents"},{"id":"2605.13625","title":"How to Interpret Agent Behavior","created":"2026-05-13","cited_by_count":10,"area":null},{"id":"2310.17769","title":"Social Contract AI: Aligning AI Assistants with Implicit Group Norms","created":"2023-10-26","cited_by_count":9,"area":null},{"id":"2410.11055","title":"Varying Shades of Wrong: Aligning LLMs with Wrong Answers Only","created":"2024-10-14","cited_by_count":9,"area":"alignment"},{"id":"2410.13826","title":"Unearthing Skill-Level Insights for Understanding Trade-Offs of Foundation Models","created":"2024-10-17","cited_by_count":9,"area":null},{"id":"2608.05179","title":"Autonomous Research Agents: A Survey of AI Scientists and the Verification Gap","created":"2026-06-29","cited_by_count":9,"area":"agents"},{"id":"2608.13331","title":"Training AI Scientists to Replicate Research","created":"2026-08-13","cited_by_count":9,"area":null},{"id":"2405.18718","title":"Efficient Model-agnostic Alignment via Bayesian Persuasion","created":"2024-05-29","cited_by_count":8,"area":"alignment"},{"id":"2505.10426","title":"Formalising Human-in-the-Loop: Computational Reductions, Failure Modes, and Legal-Moral Responsibility","created":"2025-05-15","cited_by_count":8,"area":null},{"id":"2506.06414","title":"Benchmarking Misuse Mitigation Against Covert Adversaries","created":"2025-06-06","cited_by_count":8,"area":null}],"next_cursor":"100"}