{"total":123,"rows":[{"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":"2209.00626","title":"The Alignment Problem from a Deep Learning Perspective","created":"2022-08-30","cited_by_count":258,"area":null},{"id":"2306.04488","title":"Rewarded soups: towards Pareto-optimal alignment by interpolating weights fine-tuned on diverse rewards","created":"2023-06-07","cited_by_count":243,"area":"alignment"},{"id":"2308.14752","title":"AI Deception: A Survey of Examples, Risks, and Potential Solutions","created":"2023-08-28","cited_by_count":188,"area":null},{"id":"2502.17424","title":"Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs","created":"2025-02-24","cited_by_count":172,"area":"safety"},{"id":"2302.10329","title":"Harms from Increasingly Agentic Algorithmic Systems","created":"2023-02-20","cited_by_count":135,"area":"agents"},{"id":"2210.01790","title":"Goal Misgeneralization: Why Correct Specifications Aren't Enough For Correct Goals","created":"2022-10-04","cited_by_count":126,"area":null},{"id":"2404.00282","title":"Survey on Large Language Model-Enhanced Reinforcement Learning: Concept, Taxonomy, and Methods","created":"2024-03-30","cited_by_count":116,"area":null},{"id":"2306.00937","title":"STEVE-1: A Generative Model for Text-to-Behavior in Minecraft","created":"2023-06-01","cited_by_count":97,"area":null},{"id":"2402.10877","title":"Robust agents learn causal world models","created":"2024-02-16","cited_by_count":81,"area":"wm"},{"id":"2406.04268","title":"Open-Endedness is Essential for Artificial Superhuman Intelligence","created":"2024-06-06","cited_by_count":80,"area":null},{"id":"2304.00612","title":"Eight Things to Know about Large Language Models","created":"2023-04-02","cited_by_count":73,"area":null},{"id":"2204.06601","title":"Causal Confusion and Reward Misidentification in Preference-Based Reward Learning","created":"2022-04-13","cited_by_count":53,"area":"alignment"},{"id":"2408.16984","title":"Beyond Preferences in AI Alignment","created":"2024-08-30","cited_by_count":49,"area":"alignment"},{"id":"2310.17688","title":"Managing extreme AI risks amid rapid progress","created":"2023-10-26","cited_by_count":48,"area":null},{"id":"2308.12270","title":"Language Reward Modulation for Pretraining Reinforcement Learning","created":"2023-08-23","cited_by_count":43,"area":null},{"id":"2410.04332","title":"Gradient Routing: Masking Gradients to Localize Computation in Neural Networks","created":"2024-10-06","cited_by_count":34,"area":null},{"id":"2407.18369","title":"AI Safety in Generative AI Large Language Models: A Survey","created":"2024-07-06","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":"2306.08649","title":"OCAtari: Object-Centric Atari 2600 Reinforcement Learning Environments","created":"2023-06-14","cited_by_count":30,"area":null},{"id":"2512.16912","title":"Exploration vs Exploitation: Rethinking RLVR through Clipping, Entropy, and Spurious Reward","created":"2025-12-18","cited_by_count":27,"area":"lr"},{"id":"2304.06528","title":"Power-seeking can be probable and predictive for trained agents","created":"2023-04-13","cited_by_count":25,"area":null},{"id":"2310.11971","title":"Improving Generalization of Alignment with Human Preferences through Group Invariant Learning","created":"2023-10-18","cited_by_count":25,"area":"alignment"},{"id":"2401.05821","title":"Interpretable Concept Bottlenecks to Align Reinforcement Learning Agents","created":"2024-01-11","cited_by_count":23,"area":"deeprl"},{"id":"2310.08043","title":"Understanding and Controlling a Maze-Solving Policy Network","created":"2023-10-12","cited_by_count":22,"area":null},{"id":"2403.04471","title":"The Shutdown Problem: An AI Engineering Puzzle for Decision Theorists","created":"2024-03-07","cited_by_count":22,"area":null},{"id":"2407.15421","title":"Planning in a recurrent neural network that plays Sokoban","created":"2024-07-22","cited_by_count":18,"area":null},{"id":"2503.13551","title":"Towards Hierarchical Multi-Step Reward Models for Enhanced Reasoning in Large Language Models","created":"2025-03-16","cited_by_count":18,"area":"lr"},{"id":"2302.12689","title":"GANterfactual-RL: Understanding Reinforcement Learning Agents' Strategies through Visual Counterfactual Explanations","created":"2023-02-24","cited_by_count":17,"area":null},{"id":"2308.10797","title":"Stabilizing Unsupervised Environment Design with a Learned Adversary","created":"2023-08-21","cited_by_count":17,"area":null},{"id":"2211.11030","title":"Adversarial Cheap Talk","created":"2022-11-20","cited_by_count":16,"area":null},{"id":"2504.11844","title":"Evaluating the Goal-Directedness of Large Language Models","created":"2025-04-16","cited_by_count":16,"area":null},{"id":"2406.03997","title":"HackAtari: Atari Learning Environments for Robust and Continual Reinforcement Learning","created":"2024-06-06","cited_by_count":11,"area":"deeprl"},{"id":"2505.16142","title":"RLKD: Distilling LLMs' Reasoning via Reinforcement Learning","created":"2025-05-22","cited_by_count":11,"area":"lr"},{"id":"2306.02419","title":"Bad Habits: Policy Confounding and Out-of-Trajectory Generalization in RL","created":"2023-06-04","cited_by_count":10,"area":null},{"id":"2309.10790","title":"Guide Your Agent with Adaptive Multimodal Rewards","created":"2023-09-19","cited_by_count":10,"area":"offlinerl"},{"id":"2402.10184","title":"Reward Generalization in RLHF: A Topological Perspective","created":"2024-02-15","cited_by_count":9,"area":"alignment"},{"id":"2410.23953","title":"Representative Social Choice: From Learning Theory to AI Alignment","created":"2024-10-31","cited_by_count":8,"area":"alignment"},{"id":"2406.18777","title":"Aligning Model Properties via Conformal Risk Control","created":"2024-06-26","cited_by_count":7,"area":"alignment"},{"id":"2412.04758","title":"Measuring Goal-Directedness","created":"2024-12-06","cited_by_count":7,"area":null},{"id":"2402.12907","title":"Roadmap on Incentive Compatibility for AI Alignment and Governance in Sociotechnical Systems","created":"2024-02-20","cited_by_count":6,"area":null},{"id":"2505.10597","title":"Two Minds Better Than One: Collaborative Reward Modeling for LLM Alignment","created":"2025-05-15","cited_by_count":6,"area":"alignment"},{"id":"2511.02109","title":"Deep Value Benchmark: Measuring Whether Models Generalize Deep Values or Shallow Preferences","created":"2025-11-03","cited_by_count":6,"area":"alignment"},{"id":"2512.12046","title":"Goal Reaching with Eikonal-Constrained Hierarchical Quasimetric Reinforcement Learning","created":"2025-12-12","cited_by_count":6,"area":"deeprl"},{"id":"2504.20157","title":"Toward Evaluative Thinking: Meta Policy Optimization with Evolving Reward Models","created":"2025-04-28","cited_by_count":5,"area":"lr"},{"id":"2507.12872","title":"Manipulation Attacks by Misaligned AI: Risk Analysis and Safety Case Framework","created":"2025-07-17","cited_by_count":5,"area":null},{"id":"2603.10028","title":"How to Count AIs: Individuation and Liability for AI Agents","created":"2026-02-24","cited_by_count":5,"area":null},{"id":"2605.11134","title":"Spurious Correlation Learning in Preference Optimization: Mechanisms, Consequences, and Mitigation via Tie Training","created":"2026-05-11","cited_by_count":5,"area":"alignment"},{"id":"2605.23989","title":"Towards trustworthy agentic AI: a comprehensive survey of safety, robustness, privacy, and system security","created":"2026-05-17","cited_by_count":5,"area":"agents"},{"id":"2305.07888","title":"Consistency Regularization for Domain Generalization with Logit Attribution Matching","created":"2023-05-13","cited_by_count":4,"area":null},{"id":"2407.12161","title":"Interpretability in Action: Exploratory Analysis of VPT, a Minecraft Agent","created":"2024-07-16","cited_by_count":4,"area":"agents"},{"id":"2410.14371","title":"Interpretable end-to-end Neurosymbolic Reinforcement Learning agents","created":"2024-10-18","cited_by_count":4,"area":"deeprl"},{"id":"2410.22459","title":"Predicting Future Actions of Reinforcement Learning Agents","created":"2024-10-29","cited_by_count":4,"area":"deeprl"},{"id":"2506.03755","title":"Misalignment or misuse? The AGI alignment tradeoff","created":"2025-06-04","cited_by_count":4,"area":null},{"id":"2506.05070","title":"RIVAL: Reinforcement Learning with Iterative and Adversarial Optimization for Machine Translation","created":"2025-06-05","cited_by_count":4,"area":"alignment"},{"id":"2506.10138","title":"Path Channels and Plan Extension Kernels: a Mechanistic Description of Planning in a Sokoban RNN","created":"2025-06-11","cited_by_count":4,"area":"deeprl"},{"id":"2510.21332","title":"Weak-to-Strong Generalization under Distribution Shifts","created":"2025-10-24","cited_by_count":4,"area":null},{"id":"2601.07524","title":"Stagewise Reinforcement Learning and the Geometry of the Regret Landscape","created":"2026-01-12","cited_by_count":4,"area":"deeprl"},{"id":"2604.08169","title":"Activation Steering for Aligned Open-ended Generation without Sacrificing Coherence","created":"2026-04-09","cited_by_count":4,"area":"safety"},{"id":"2605.15960","title":"Imperfect World Models are Exploitable","created":"2026-05-15","cited_by_count":4,"area":"wm"},{"id":"2606.28430","title":"Building to the Test: Coding Agents Deliver What You Check, Not What You Requested","created":"2026-06-26","cited_by_count":4,"area":"agents"},{"id":"2502.04512","title":"Safety Must Precede the Deployment of Open-Ended AI","created":"2025-02-06","cited_by_count":3,"area":"deeprl"},{"id":"2504.07257","title":"Better Decisions through the Right Causal World Model","created":"2025-04-09","cited_by_count":3,"area":"wm"},{"id":"2604.01346","title":"Safety, Security, and Cognitive Risks in World Models","created":"2026-04-01","cited_by_count":3,"area":"wm"},{"id":"2606.30252","title":"Inoculation Adapters: Improved Selective Generalization of Capabilities with Fewer Surprising Backdoors","created":"2026-06-29","cited_by_count":3,"area":null},{"id":"2304.11104","title":"Approximate Shielding of Atari Agents for Safe Exploration","created":"2023-04-21","cited_by_count":2,"area":null},{"id":"2312.17168","title":"Can Active Sampling Reduce Causal Confusion in Offline Reinforcement Learning?","created":"2023-12-28","cited_by_count":2,"area":"offlinerl"},{"id":"2407.00805","title":"Towards Shutdownable Agents via Stochastic Choice","created":"2024-06-30","cited_by_count":2,"area":null},{"id":"2505.02313","title":"What Is AI Safety? What Do We Want It to Be?","created":"2025-05-05","cited_by_count":2,"area":null},{"id":"2508.19562","title":"Democracy-in-Silico: Institutional Design as Alignment in AI-Governed Polities","created":"2025-08-27","cited_by_count":2,"area":null},{"id":"2602.01942","title":"Human Society-Inspired Approaches to Agentic AI Security: The 4C Framework","created":"2026-02-02","cited_by_count":2,"area":"agents"},{"id":"2602.02212","title":"MAIN-VLA: Modeling Abstraction of Intention and eNvironment for Vision-Language-Action Models","created":"2026-02-02","cited_by_count":2,"area":"vla"},{"id":"2603.09400","title":"Reward Prediction with Factorized World States","created":"2026-03-10","cited_by_count":2,"area":"alignment"},{"id":"2604.12086","title":"Robust Optimization for Mitigating Reward Hacking with Correlated Proxies","created":"2026-04-13","cited_by_count":2,"area":"alignment"},{"id":"2605.01643","title":"AI Alignment via Incentives and Correction","created":"2026-05-02","cited_by_count":2,"area":null},{"id":"2306.10999","title":"Concept Extrapolation: A Conceptual Primer","created":"2023-06-19","cited_by_count":1,"area":null},{"id":"2309.16166","title":"CoinRun: Solving Goal Misgeneralisation","created":"2023-09-28","cited_by_count":1,"area":null},{"id":"2312.03762","title":"Colour versus Shape Goal Misgeneralization in Reinforcement Learning: A Case Study","created":"2023-12-05","cited_by_count":1,"area":null},{"id":"2404.12150","title":"Aligning language models with human preferences","created":"2024-04-18","cited_by_count":1,"area":"alignment"},{"id":"2407.16025","title":"Exploring and Addressing Reward Confusion in Offline Preference Learning","created":"2024-07-22","cited_by_count":1,"area":"alignment"},{"id":"2408.08995","title":"On the Undecidability of Artificial Intelligence Alignment: Machines that Halt","created":"2024-08-16","cited_by_count":1,"area":"alignment"},{"id":"2410.07096","title":"Rejecting Hallucinated State Targets during Planning","created":"2024-10-09","cited_by_count":1,"area":null},{"id":"2504.15125","title":"Contemplative Artificial Intelligence","created":"2025-04-21","cited_by_count":1,"area":null},{"id":"2509.15932","title":"The Alignment Bottleneck","created":"2025-09-19","cited_by_count":1,"area":null},{"id":"2510.11235","title":"AI Alignment Strategies from a Risk Perspective: Independent Safety Mechanisms or Shared Failures?","created":"2025-10-13","cited_by_count":1,"area":null},{"id":"2601.17344","title":"The Shadow Self: Intrinsic Value Misalignment in Large Language Model Agents","created":"2026-01-24","cited_by_count":1,"area":"agents"},{"id":"2601.21350","title":"Factored Causal Representation Learning for Robust Reward Modeling in RLHF","created":"2026-01-29","cited_by_count":1,"area":"alignment"},{"id":"2602.01699","title":"Mitigating loss of control in advanced AI systems through instrumental goal trajectories","created":"2026-02-02","cited_by_count":1,"area":null},{"id":"2604.19001","title":"HarmThoughts: A Benchmark for Fine-Grained Harmful Behavior Detection in Reasoning Traces","created":"2026-04-21","cited_by_count":1,"area":"lr"},{"id":"2604.26360","title":"Uncertainty-Aware Reward Discounting for Mitigating Reward Hacking","created":"2026-04-29","cited_by_count":1,"area":"alignment"},{"id":"2607.22188","title":"Draining the Energy Commons: Self-Defeating Over-Appropriation as a Coordination Failure in Agentic LLM Collectives","created":"2026-07-24","cited_by_count":1,"area":"agents"},{"id":"2607.25379","title":"Cyber-Capable AI Agents: Vulnerabilities, Evaluation Containment, and Defensive Response","created":"2026-07-28","cited_by_count":1,"area":"agents"},{"id":"2306.13000","title":"Apolitical Intelligence? Auditing Delphi's responses on controversial political issues in the US","created":"2023-06-22","cited_by_count":0,"area":null},{"id":"2310.03494","title":"How the level sampling process impacts zero-shot generalisation in deep reinforcement learning","created":"2023-10-05","cited_by_count":0,"area":null},{"id":"2310.05871","title":"Dynamic value alignment through preference aggregation of multiple objectives","created":"2023-10-09","cited_by_count":0,"area":null},{"id":"2311.12004","title":"Risk-averse Batch Active Inverse Reward Design","created":"2023-11-20","cited_by_count":0,"area":null},{"id":"2402.09734","title":"Agents Need Not Know Their Purpose","created":"2024-02-15","cited_by_count":0,"area":null},{"id":"2404.10906","title":"Towards a Research Community in Interpretable Reinforcement Learning: the InterpPol Workshop","created":"2024-04-16","cited_by_count":0,"area":"deeprl"},{"id":"2411.00867","title":"Mechanistic Interpretability of Reinforcement Learning Agents","created":"2024-10-30","cited_by_count":0,"area":"mechinterp"},{"id":"2411.04976","title":"Noisy Zero-Shot Coordination: Breaking The Common Knowledge Assumption In Zero-Shot Coordination Games","created":"2024-11-07","cited_by_count":0,"area":"deeprl"}],"next_cursor":"100"}