{"total":96,"rows":[{"id":"2405.14860","title":"Not All Language Model Features Are One-Dimensionally Linear","created":"2024-05-23","cited_by_count":212,"area":null},{"id":"2504.15585","title":"A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deployment","created":"2025-04-22","cited_by_count":116,"area":"lr"},{"id":"2411.12990","title":"BetterBench: Assessing AI Benchmarks, Uncovering Issues, and Establishing Best Practices","created":"2024-11-20","cited_by_count":114,"area":null},{"id":"2412.17686","title":"Large Language Model Safety: A Holistic Survey","created":"2024-12-23","cited_by_count":59,"area":"safety"},{"id":"2407.15017","title":"Knowledge Mechanisms in Large Language Models: A Survey and Perspective","created":"2024-07-22","cited_by_count":51,"area":null},{"id":"2501.10114","title":"Infrastructure for AI Agents","created":"2025-01-17","cited_by_count":41,"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":"2502.11355","title":"Nuclear Deployed: Analyzing Catastrophic Risks in Decision-making of Autonomous LLM Agents","created":"2025-02-17","cited_by_count":27,"area":"lr"},{"id":"2410.09713","title":"Agentic Information Retrieval","created":"2024-10-13","cited_by_count":21,"area":"agents"},{"id":"2602.06176","title":"Large Language Model Reasoning Failures","created":"2026-02-05","cited_by_count":21,"area":"lr"},{"id":"2405.15454","title":"LiSeCo: Linear Semantic Control for Language Generation","created":"2024-05-24","cited_by_count":20,"area":null},{"id":"2406.11779","title":"Compact Proofs of Model Performance via Mechanistic Interpretability","created":"2024-06-17","cited_by_count":18,"area":"mechinterp"},{"id":"2502.20432","title":"LLM Strategic Reasoning: Agentic Study through Behavioral Game Theory","created":"2025-02-27","cited_by_count":17,"area":"lr"},{"id":"2506.01622","title":"General agents contain world models","created":"2025-06-02","cited_by_count":17,"area":"wm"},{"id":"2408.02205","title":"Swiss Cheese Model for AI Safety: A Taxonomy and Reference Architecture for Multi-Layered Guardrails of Foundation Model Based Agents","created":"2024-08-05","cited_by_count":14,"area":"agents"},{"id":"2410.18819","title":"From Imitation to Introspection: Probing Self-Consciousness in Language Models","created":"2024-10-24","cited_by_count":14,"area":null},{"id":"2602.08235","title":"When Benign Inputs Lead to Severe Harms: Eliciting Unsafe Unintended Behaviors of Computer-Use Agents","created":"2026-02-09","cited_by_count":14,"area":"agents"},{"id":"2408.05284","title":"Can a Bayesian Oracle Prevent Harm from an Agent?","created":"2024-08-09","cited_by_count":13,"area":null},{"id":"2505.24445","title":"Learning Safety Constraints for Large Language Models","created":"2025-05-30","cited_by_count":13,"area":"safety"},{"id":"2411.02478","title":"Imagining and building wise machines: The centrality of AI metacognition","created":"2024-11-04","cited_by_count":12,"area":null},{"id":"2410.16665","title":"SafetyAnalyst: Interpretable, Transparent, and Steerable Safety Moderation for AI Behavior","created":"2024-10-22","cited_by_count":10,"area":null},{"id":"2506.20702","title":"The Singapore Consensus on Global AI Safety Research Priorities","created":"2025-06-25","cited_by_count":10,"area":null},{"id":"2502.06656","title":"A Frontier AI Risk Management Framework: Bridging the Gap Between Current AI Practices and Established Risk Management","created":"2025-02-10","cited_by_count":8,"area":null},{"id":"2506.17434","title":"Resource Rational Contractualism Should Guide AI Alignment","created":"2025-06-20","cited_by_count":8,"area":null},{"id":"2501.18426","title":"Guaranteed prediction sets for functional surrogate models","created":"2025-01-30","cited_by_count":7,"area":null},{"id":"2502.14445","title":"PredictaBoard: Benchmarking LLM Score Predictability","created":"2025-02-20","cited_by_count":7,"area":null},{"id":"2506.11083","title":"RedDebate: Safer Responses Through Multi-Agent Red Teaming Debates","created":"2025-06-04","cited_by_count":7,"area":"agents"},{"id":"2510.06296","title":"VeriEquivBench: An Equivalence Score for Ground-Truth-Free Evaluation of Formally Verifiable Code","created":"2025-10-07","cited_by_count":7,"area":null},{"id":"2601.19612","title":"Safe Exploration via Policy Priors","created":"2026-01-27","cited_by_count":6,"area":"deeprl"},{"id":"2507.15851","title":"The Other Mind: How Language Models Exhibit Human Temporal Cognition","created":"2025-07-21","cited_by_count":5,"area":null},{"id":"2601.10520","title":"Breaking Up with Normatively Monolithic Agency with GRACE: A Reason-Based Neuro-Symbolic Architecture for Safe and Ethical AI Alignment","created":"2026-01-15","cited_by_count":5,"area":"lr"},{"id":"2605.10310","title":"Positive Alignment: Artificial Intelligence for Human Flourishing","created":"2026-05-11","cited_by_count":5,"area":null},{"id":"2406.00826","title":"Policy Verification in Stochastic Dynamical Systems Using Logarithmic Neural Certificates","created":"2024-06-02","cited_by_count":4,"area":null},{"id":"2406.02619","title":"Unelicitable Backdoors in Language Models via Cryptographic Transformer Circuits","created":"2024-06-03","cited_by_count":4,"area":"safety"},{"id":"2407.12999","title":"Securing the Future of GenAI: Policy and Technology","created":"2024-05-21","cited_by_count":4,"area":null},{"id":"2410.08255","title":"Investigating Representation Universality: Case Study on Genealogical Representations","created":"2024-10-10","cited_by_count":4,"area":null},{"id":"2502.03467","title":"Where AI Assurance Might Go Wrong: Initial lessons from engineering of critical systems","created":"2025-01-07","cited_by_count":4,"area":null},{"id":"2505.15811","title":"On the creation of narrow AI: hierarchy and nonlocality of neural network skills","created":"2025-05-21","cited_by_count":4,"area":null},{"id":"2512.02193","title":"From monoliths to modules: Decomposing transducers for efficient world modelling","created":"2025-12-01","cited_by_count":4,"area":"wm"},{"id":"2604.19459","title":"Do LLMs Game Formalization? Evaluating Faithfulness in Logical Reasoning","created":"2026-04-21","cited_by_count":4,"area":"lr"},{"id":"2604.27292","title":"The Two Boundaries: Why Behavioral AI Governance Fails Structurally","created":"2026-04-30","cited_by_count":4,"area":null},{"id":"2607.22868","title":"What Can Be Enforced? A Theory of Certified Runtime Safety for Tool-Using Agents","created":"2026-07-24","cited_by_count":4,"area":"safety"},{"id":"2409.03793","title":"Safeguarding AI Agents: Developing and Analyzing Safety Architectures","created":"2024-09-03","cited_by_count":3,"area":"agents"},{"id":"2504.04608","title":"AI in a vat: Fundamental limits of efficient world modelling for agent sandboxing and interpretability","created":"2025-04-06","cited_by_count":3,"area":"wm"},{"id":"2506.02923","title":"The Limits of Predicting Agents from Behaviour","created":"2025-06-03","cited_by_count":3,"area":null},{"id":"2506.17846","title":"Out of Control -- Why Alignment Needs Formal Control Theory (and an Alignment Control Stack)","created":"2025-06-21","cited_by_count":3,"area":null},{"id":"2507.03525","title":"Limits of Safe AI Deployment: Differentiating Oversight and Control","created":"2025-07-04","cited_by_count":3,"area":null},{"id":"2605.03034","title":"Stable Agentic Control: Tool-Mediated LLM Architecture for Autonomous Cyber Defense","created":"2026-05-04","cited_by_count":3,"area":"agents"},{"id":"2501.10476","title":"Revisiting Rogers' Paradox in the Context of Human-AI Interaction","created":"2025-01-16","cited_by_count":2,"area":null},{"id":"2502.04249","title":"Free Energy Risk Metrics for Systemically Safe AI: Gatekeeping Multi-Agent Study","created":"2025-02-06","cited_by_count":2,"area":"wm"},{"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":"2511.11816","title":"Do LLMs Really Struggle at NL-FOL Translation? Revealing their Strengths via a Novel Benchmarking Strategy","created":"2025-11-14","cited_by_count":2,"area":null},{"id":"2604.11174","title":"EmbodiedGovBench: A Benchmark for Governance, Recovery, and Upgrade Safety in Embodied Agent Systems","created":"2026-04-13","cited_by_count":2,"area":"agents"},{"id":"2410.18114","title":"Bridging Today and the Future of Humanity: AI Safety in 2024 and Beyond","created":"2024-10-09","cited_by_count":1,"area":null},{"id":"2412.15433","title":"Quantifying detection rates for dangerous capabilities: a theoretical model of dangerous capability evaluations","created":"2024-12-19","cited_by_count":1,"area":null},{"id":"2501.11183","title":"Can Safety Fine-Tuning Be More Principled? Lessons Learned from Cybersecurity","created":"2025-01-19","cited_by_count":1,"area":"safety"},{"id":"2505.14689","title":"Follow the STARs: Dynamic $\\omega$-Regular Shielding of Learned Policies","created":"2025-04-11","cited_by_count":1,"area":null},{"id":"2506.16971","title":"Formal Control for Uncertain Systems via Contract-Based Probabilistic Surrogates (Extended Version)","created":"2025-06-20","cited_by_count":1,"area":null},{"id":"2509.25843","title":"ASGuard: Activation-Scaling Guard to Mitigate Targeted Jailbreaking Attack","created":"2025-09-30","cited_by_count":1,"area":"safety"},{"id":"2510.13727","title":"From Refusal to Recovery: A Control-Theoretic Approach to Generative AI Guardrails","created":"2025-10-15","cited_by_count":1,"area":"safety"},{"id":"2512.20755","title":"Bridging Efficiency and Safety: Formal Verification of Neural Networks with Early Exits","created":"2025-12-23","cited_by_count":1,"area":"inference"},{"id":"2601.22356","title":"PoSafeNet: Safe Learning with Poset-Structured Neural Nets","created":"2026-01-29","cited_by_count":1,"area":null},{"id":"2602.23997","title":"Foundation World Models for Agents that Learn, Verify, and Adapt Reliably Beyond Static Environments","created":"2026-02-27","cited_by_count":1,"area":"wm"},{"id":"2603.08760","title":"Clear, Compelling Arguments: Rethinking the Foundations of Frontier AI Safety Cases","created":"2026-03-08","cited_by_count":1,"area":null},{"id":"2603.12183","title":"Proof-Carrying Materials: Falsifiable Safety Certificates for Machine-Learned Interatomic Potentials","created":"2026-03-12","cited_by_count":1,"area":null},{"id":"2604.13609","title":"Golden Handcuffs make safer AI agents","created":"2026-04-15","cited_by_count":1,"area":null},{"id":"2605.01030","title":"Effect-Transparent Governance for AI Workflow Architectures: Semantic Preservation, Expressive Minimality, and Decidability Boundaries","created":"2026-05-01","cited_by_count":1,"area":null},{"id":"2605.01032","title":"Algebraic Semantics of Governed Execution: Monoidal Categories, Effect Algebras, and Coterminous Boundaries","created":"2026-05-01","cited_by_count":1,"area":null},{"id":"2606.02837","title":"Fixing FOLIO and MALLS: Verified Annotations and an LLM-assisted Framework to Focus Human Relabeling","created":"2026-06-01","cited_by_count":1,"area":null},{"id":"2606.04037","title":"Toward Pre-Deployment Assurance for Enterprise AI Agents: Ontology-Grounded Simulation and Trust Certification","created":"2026-06-02","cited_by_count":1,"area":"agents"},{"id":"2609.00092","title":"Safin-1: Safety from Within through Memory-Native State Evolution","created":"2026-08-31","cited_by_count":1,"area":null},{"id":"2407.11001","title":"Generative AI Systems: A Systems-based Perspective on Generative AI","created":"2024-06-25","cited_by_count":0,"area":null},{"id":"2407.11360","title":"Thorns and Algorithms: Navigating Generative AI Challenges Inspired by Giraffes and Acacias","created":"2024-07-16","cited_by_count":0,"area":null},{"id":"2408.04385","title":"Non-maximizing policies that fulfill multi-criterion aspirations in expectation","created":"2024-08-08","cited_by_count":0,"area":"deeprl"},{"id":"2412.16238","title":"Algebraic Evaluation Theorems","created":"2024-12-19","cited_by_count":0,"area":null},{"id":"2503.00563","title":"A Guide to Failure in Machine Learning: Reliability and Robustness from Foundations to Practice","created":"2025-03-01","cited_by_count":0,"area":null},{"id":"2505.20110","title":"Beyond the Proxy: Trajectory-Distilled Guidance for Offline GFlowNet Training","created":"2025-05-26","cited_by_count":0,"area":null},{"id":"2506.05171","title":"Towards provable probabilistic safety for scalable embodied AI systems","created":"2025-06-05","cited_by_count":0,"area":null},{"id":"2506.22492","title":"Report on NSF Workshop on Science of Safe AI","created":"2025-06-24","cited_by_count":0,"area":null},{"id":"2602.12851","title":"Chimera: Neuro-Symbolic Attention Primitives for Trustworthy Dataplane Intelligence","created":"2026-02-13","cited_by_count":0,"area":null},{"id":"2602.17676","title":"Epistemic Traps: Rational Misalignment Driven by Model Misspecification","created":"2026-01-27","cited_by_count":0,"area":null},{"id":"2602.20102","title":"BarrierSteer: LLM Safety via Learning Barrier Steering","created":"2026-02-23","cited_by_count":0,"area":"safety"},{"id":"2603.25531","title":"Synchronous Signal Temporal Logic for Decidable Verification of Cyber-Physical Systems","created":"2026-03-26","cited_by_count":0,"area":null},{"id":"2605.12963","title":"Sustaining AI safety: Control-theoretic external impossibility, intrinsic necessity, and structural requirements","created":"2026-05-13","cited_by_count":0,"area":null},{"id":"2605.26940","title":"Accountable Human-AI Deliberation with LLMs: Scaling Collective Intelligence through Symbiotic Scaffolding","created":"2026-05-26","cited_by_count":0,"area":null},{"id":"2606.01008","title":"FVSpec: Real-World Property-Based Tests as Lean Challenges","created":"2026-05-31","cited_by_count":0,"area":null},{"id":"2606.04823","title":"R-APS: Compositional Reasoning and In-Context Meta-Learning for Constrained Design via Reflective Adversarial Pareto Search","created":"2026-06-03","cited_by_count":0,"area":"lr"},{"id":"2606.09940","title":"Interactions Between Crosscoder Features: A Compact Proofs Perspective","created":"2026-06-08","cited_by_count":0,"area":"mechinterp"},{"id":"2607.16224","title":"International Agreements to Limit Frontier AI: Objectives and Exit","created":"2026-06-19","cited_by_count":0,"area":null},{"id":"2607.28954","title":"Mining Verdict Boundaries for Neural Network Verification","created":"2026-07-31","cited_by_count":0,"area":null},{"id":"2608.11274","title":"Agent Safety Should Be a Runtime Contract","created":"2026-08-11","cited_by_count":0,"area":"agents"},{"id":"2609.07627","title":"Norms at a Price: Why RL-Based Alignment Can Promise Conditional Compliance at Best","created":"2026-09-07","cited_by_count":0,"area":null},{"id":"2609.22712","title":"Trustworthy Agentic AI: Failure Modes, Mitigation Strategies, and a Lifecycle Framework for Autonomous LLM Systems","created":"2026-09-19","cited_by_count":0,"area":"agents"},{"id":"2610.09159","title":"SpecGuard: Proving a Task Is Broken Before the Agent Cheats","created":"2026-10-06","cited_by_count":0,"area":"agents"},{"id":"2610.10285","title":"AI Safety Considerations for Agents With Limited Time to Act","created":"2026-10-07","cited_by_count":0,"area":null},{"id":"2610.12030","title":"Natural Language to First-Order Logic LLM-based Autoformalization","created":"2026-10-08","cited_by_count":0,"area":null}],"next_cursor":null}