{"total":138,"rows":[{"id":"2609.04172","title":"Rethinking On-Policy Distillation of Large Language Models II: One Training Example","created":"2026-09-03","cited_by_count":24,"area":"distill"},{"id":"2608.09867","title":"Stealing Reasoning Traces from Proprietary LLM APIs","created":"2026-08-10","cited_by_count":9,"area":"lr"},{"id":"2608.18271","title":"Rethinking Privileged Information in On-Policy Self-Distillation","created":"2026-08-18","cited_by_count":9,"area":"distill"},{"id":"2609.04565","title":"Extremely Sparse Supervision Incentivizes Reasoning Ability","created":"2026-09-03","cited_by_count":5,"area":"lr"},{"id":"2608.17310","title":"Agentic ESOpt: Fine-Tuning Long-Horizon LLM Agents with Minimal GPU Requirements","created":"2026-08-18","cited_by_count":3,"area":"agents"},{"id":"2609.08977","title":"Multimodal Duplex Interaction Agent","created":"2026-09-08","cited_by_count":3,"area":null},{"id":"2609.20784","title":"RetireOPD: Self-Retiring On-Policy Distillation for Agentic Reinforcement Learning","created":"2026-09-17","cited_by_count":3,"area":"agents"},{"id":"2608.06197","title":"EnvACE: Internalizing Environment Dynamics via World Rehearsal for Agentic Reinforcement Learning","created":"2026-08-06","cited_by_count":2,"area":"wm"},{"id":"2609.04010","title":"Unlocking Lossless Speedups in LLMs via Discrete Diffusion","created":"2026-09-03","cited_by_count":2,"area":"diffusion"},{"id":"2609.06498","title":"DFlow: Enabling Verifier Information Flow in Block Diffusion Speculative Decoding","created":"2026-09-06","cited_by_count":2,"area":"inference"},{"id":"2609.08798","title":"Eliciting Weak-to-Strong Generalization with On-Policy Reverse Distillation","created":"2026-09-08","cited_by_count":2,"area":"distill"},{"id":"2609.18708","title":"Rethinking Critic Learning in PPO: Understanding and Mitigating Value Flattening","created":"2026-09-16","cited_by_count":2,"area":"lr"},{"id":"2609.19101","title":"Monitoring and Discovering Reward Hacking with Internal Representations during LLM Evaluations","created":"2026-09-16","cited_by_count":2,"area":"alignment"},{"id":"2610.07557","title":"CheckerBench: Can Long-Horizon Agents Synthesize Static-Analysis Checkers?","created":"2026-10-06","cited_by_count":2,"area":null},{"id":"2610.13778","title":"ReplaySSM: Cache SSM Inputs, Not State","created":"2026-10-09","cited_by_count":2,"area":"inference"},{"id":"2608.02441","title":"Agentic Commerce World: An Auditable and Verifiable Environment for Vibe Commerce","created":"2026-08-03","cited_by_count":1,"area":"agents"},{"id":"2608.11152","title":"Scheduling Mixed RL Rollouts Beyond Prefix Locality","created":"2026-08-11","cited_by_count":1,"area":null},{"id":"2608.13482","title":"Synthetic Persona Pretraining: Alignment from Token Zero","created":"2026-08-13","cited_by_count":1,"area":null},{"id":"2608.30505","title":"Tensor Methods for Language Models: From Token Representation to Training, Adaptation, Inference, Compression, and Interpretability","created":"2026-08-31","cited_by_count":1,"area":null},{"id":"2609.02737","title":"Language Models Can Control Their Own Attention","created":"2026-09-02","cited_by_count":1,"area":"inference"},{"id":"2609.10494","title":"IBIB: A Protocol for Measuring Enterprise AI Systems by Serving Route, Not Model Identifier","created":"2026-09-09","cited_by_count":1,"area":null},{"id":"2609.11319","title":"Magenta: Closing the Loop Between Mathematical Reasoning and Lean Verification","created":"2026-09-10","cited_by_count":1,"area":"lr"},{"id":"2609.24526","title":"ME-VLM: A Unified VLM for Embodied Cognition and Agent Coordination","created":"2026-09-21","cited_by_count":1,"area":"agents"},{"id":"2609.29204","title":"AdaHVLA: Adaptive Harnesses for Long-Horizon Vision-Language-Action Execution","created":"2026-09-24","cited_by_count":1,"area":"vla"},{"id":"2609.32069","title":"Find Something You Can't Do: Agentic Real-World Reinforcement Learning for Self-Improving VLA Models","created":"2026-09-25","cited_by_count":1,"area":"vla"},{"id":"2609.32534","title":"DepthBench: Measuring How Residual Connections Enable More Computational Depth","created":"2026-09-26","cited_by_count":1,"area":null},{"id":"2609.32577","title":"Groupwise Agentic Grading and Advantage Redistribution for Code Agent RL","created":"2026-09-26","cited_by_count":1,"area":"agents"},{"id":"2609.35751","title":"How to Loop MoE: Flatten the Experts, Untie the Attention","created":"2026-09-28","cited_by_count":1,"area":"moe"},{"id":"2610.04184","title":"EvalResearchBench: Can AI Agents Design Their Own Evaluations?","created":"2026-10-03","cited_by_count":1,"area":null},{"id":"2608.13675","title":"From BERT to Frontier Agents: Eight Years of Language-Model Progress, the Collapse of the Capability-Cost Curve, and the Rise of Task-Targeted Models","created":"2026-08-13","cited_by_count":0,"area":null},{"id":"2608.15012","title":"SysEvolve: An AI-native, safe, autonomous adversarial attack-defense co-evolutionary system","created":"2026-08-15","cited_by_count":0,"area":null},{"id":"2608.16760","title":"On the Principles Behind Neural Network Optimizers","created":"2026-08-17","cited_by_count":0,"area":null},{"id":"2608.26052","title":"How Much Rank Does LoRA Need? Rank-Error Bounds for Transformer Attention","created":"2026-08-26","cited_by_count":0,"area":null},{"id":"2608.30567","title":"TuringLLM: Efficiently Scaling Foundation Models Toward Physical AI","created":"2026-08-31","cited_by_count":0,"area":"moe"},{"id":"2609.00264","title":"The Answer Is Not the Argument","created":"2026-08-31","cited_by_count":0,"area":"lr"},{"id":"2609.00791","title":"Instella-MoE Technical Report","created":"2026-09-01","cited_by_count":0,"area":"moe"},{"id":"2609.03501","title":"Towards a Statistical Understanding of Mixture-of-Experts","created":"2026-09-03","cited_by_count":0,"area":"moe"},{"id":"2609.03773","title":"RealCADBench: Benchmarking Parametric CAD Modeling from Industrial Design Intents","created":"2026-09-03","cited_by_count":0,"area":null},{"id":"2609.06245","title":"VDiff-Bench: A Challenging Benchmark for Fine-Grained Image Difference Identification","created":"2026-09-05","cited_by_count":0,"area":"mllm"},{"id":"2609.06970","title":"CantoneseLLM v2: Reasoning in a Low-Resource Language","created":"2026-09-07","cited_by_count":0,"area":"lr"},{"id":"2609.07036","title":"Aha-Flow Distillation: Flow Markers Matter in LLM Reasoning","created":"2026-09-07","cited_by_count":0,"area":"lr"},{"id":"2609.07784","title":"xDailyBench: Benchmarking LLMs on Professional Consultation for Real-Life Problems","created":"2026-09-07","cited_by_count":0,"area":null},{"id":"2609.08418","title":"Feyospace-v1: How the Cyber Mercury Seven Trained Frontier Cyber Models","created":"2026-09-08","cited_by_count":0,"area":null},{"id":"2609.08690","title":"Hyperparameter Scaling Laws Across MoE Sparsity","created":"2026-09-08","cited_by_count":0,"area":"moe"},{"id":"2609.09113","title":"SAEScientist-Bench: Can AI Agents Conduct Autonomous SAE Interpretability Research?","created":"2026-09-08","cited_by_count":0,"area":"mechinterp"},{"id":"2609.12606","title":"Beyond Generation and Accuracy: Diagnosing and Enhancing Visual Chain-of-Thought for Geometry Problem Solving","created":"2026-09-11","cited_by_count":0,"area":"lr"},{"id":"2609.13356","title":"ZGCM-1: A Fully Open and Extremely Efficient Foundation Model for Math and Agentic Search","created":"2026-09-11","cited_by_count":0,"area":"agents"},{"id":"2609.15545","title":"The Token Before the Value Is the Key: How Hybrid Architectures Organize Induction Circuits","created":"2026-09-14","cited_by_count":0,"area":null},{"id":"2609.22247","title":"CHART: A Harness-Rotation Curriculum for Harness-Robust Search Agents","created":"2026-09-05","cited_by_count":0,"area":null},{"id":"2609.23578","title":"AR-WAM: A Visual-Conditioned Agent-Ready World Action Model for Robotic Manipulation","created":"2026-09-20","cited_by_count":0,"area":"wm"},{"id":"2609.24202","title":"Opinion Leader Dynamics: How Sparse Attention Shapes Token Clustering","created":"2026-09-21","cited_by_count":0,"area":null},{"id":"2609.25192","title":"FinFIRST: Benchmarking Search Agents for Financial Information Retrieval, Sourcing and Traceability","created":"2026-09-21","cited_by_count":0,"area":"agents"},{"id":"2609.25442","title":"WeightBridge: An Efficient Weight Transfer Library for Reinforcement Learning","created":"2026-09-21","cited_by_count":0,"area":"lr"},{"id":"2609.26777","title":"SWE-Serve: Benchmarking Agentic Engineering For Production Inference Serving","created":"2026-09-22","cited_by_count":0,"area":"agents"},{"id":"2609.27603","title":"When Context Misleads: In-context Learning with Jurisdiction in Large Language Models","created":"2026-09-23","cited_by_count":0,"area":null},{"id":"2609.29773","title":"Breaking the Environment Wall: Evolving LLM Agent Environments for Recursive Self-Improvement","created":"2026-09-24","cited_by_count":0,"area":"agents"},{"id":"2609.31382","title":"Highlight-Then-Summarize: Learning to Compress Evidence for Long-Context Understanding","created":"2026-09-25","cited_by_count":0,"area":null},{"id":"2609.32493","title":"SoFT: Soft Targets for Generalizable LLM Fine-Tuning","created":"2026-09-26","cited_by_count":0,"area":"distill"},{"id":"2609.32533","title":"LLMAdBench: A Human Preference Benchmark for Advertising in LLM Responses","created":"2026-09-26","cited_by_count":0,"area":null},{"id":"2609.32679","title":"The GUI Is Not the State: Diagnosing State Aliasing in GUI World Models","created":"2026-09-26","cited_by_count":0,"area":"wm"},{"id":"2609.32763","title":"Mandela-Bench: Multimodal Models Remember Canonical Images Instead of Seeing Them","created":"2026-09-26","cited_by_count":0,"area":null},{"id":"2609.32913","title":"Allspark: Weak to Strong Transfer via Alternating Chain of Thought","created":"2026-09-26","cited_by_count":0,"area":"lr"},{"id":"2609.33047","title":"Cost-free Spectral Estimation for Adaptive Newton--Schulz in Matrix Optimizers","created":"2026-09-27","cited_by_count":0,"area":null},{"id":"2609.33465","title":"GSM: Efficient Language Modeling with Shared Global State","created":"2026-09-27","cited_by_count":0,"area":null},{"id":"2609.33791","title":"Do We Really Need KL Divergence for On-Policy Distillation of Large Language Models?","created":"2026-09-27","cited_by_count":0,"area":"distill"},{"id":"2609.33807","title":"CodeActionBench: Evaluating Agentic Code-as-Policy for Embodied Manipulation","created":"2026-09-27","cited_by_count":0,"area":"agents"},{"id":"2609.34344","title":"Learning to Steer, Steering to See: Unveiling the Geometry of RLVR in Large Language Models via Trainable Vectors","created":"2026-09-28","cited_by_count":0,"area":"lr"},{"id":"2609.34492","title":"PowerBench: A Benchmark for Agentic Retrieval and Reasoning in Power Systems","created":"2026-09-28","cited_by_count":0,"area":"lr"},{"id":"2609.34605","title":"PMOPD: Task Ordering, Cycling, and Parameter-Update Subspace Protection in Multi-Teacher On-Policy Distillation","created":"2026-09-28","cited_by_count":0,"area":"distill"},{"id":"2609.34745","title":"No Pain, More Gain: Iterative Merging for Effective Multi-Teacher On-Policy Distillation","created":"2026-09-28","cited_by_count":0,"area":"distill"},{"id":"2609.34838","title":"DivOPD: Spread Wide, Look Close for Asynchronous On-Policy Distillation of Multi-turn Agents","created":"2026-09-28","cited_by_count":0,"area":"distill"},{"id":"2609.34863","title":"Revisit to Segment: Working Memory Distillation for Reasoning Segmentation","created":"2026-09-28","cited_by_count":0,"area":"lr"},{"id":"2609.35263","title":"WavePP: High-Throughput Pipeline Parallel LLM Prefill under Prefix Reuse","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2609.35561","title":"RSI-Master: Structuring Experiments to Guide Autonomous Model Improvement","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2609.35701","title":"MeqMuon: Matrix-Equilibrating Muon for LLM Pretraining","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2609.35909","title":"Cheap to Hypothesize, Costly to Verify: The Defense Surface of Agentic Vulnerability Discovery","created":"2026-09-28","cited_by_count":0,"area":"agents"},{"id":"2609.35912","title":"MMSkillRisk: Can Agents Stay Safe When Multimodal Skills Become Traps?","created":"2026-09-28","cited_by_count":0,"area":"agents"},{"id":"2609.36246","title":"Learning from Teacher Continuations at Student States","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2609.36705","title":"JudgeProfile: Understanding and Steering Subjectivity in LLM Judges","created":"2026-09-29","cited_by_count":0,"area":null},{"id":"2609.36730","title":"Can Agents Design Libraries for Agents?","created":"2026-09-29","cited_by_count":0,"area":null},{"id":"2609.37852","title":"Delta-Matching: Closing the Final Gap of Native 8-bit Training for LLMs","created":"2026-09-29","cited_by_count":0,"area":null},{"id":"2609.38335","title":"E2E-SWE: Benchmarking LLMs on Building Working Codebases from Scratch","created":"2026-09-29","cited_by_count":0,"area":"agents"},{"id":"2609.38891","title":"Consistent Plan-Act for Long-Horizon Agentic Tasks","created":"2026-09-30","cited_by_count":0,"area":"agents"},{"id":"2609.39118","title":"Diagnosing On-Policy Self-Distillation for Reasoning Language Models","created":"2026-09-30","cited_by_count":0,"area":"lr"},{"id":"2609.39137","title":"ID Balancing: Stable Training of Extremely Sparse MoE via PID-Based Load Control","created":"2026-09-30","cited_by_count":0,"area":"moe"},{"id":"2609.39325","title":"WorkGenesis: Building the Worlds That Teach Agents to Work","created":"2026-09-30","cited_by_count":0,"area":"agents"},{"id":"2609.39371","title":"EHR-RobustGym: Benchmarking and Training Agents for Robust Clinical Reasoning","created":"2026-09-30","cited_by_count":0,"area":"lr"},{"id":"2609.39490","title":"OmniReasoning: Pushing the Limits of Audio-Visual Joint Reasoning","created":"2026-09-30","cited_by_count":0,"area":"distill"},{"id":"2609.39600","title":"GroundAnything: Reconciling Parallel Decoding with Precise Visual Grounding at Flash Speed","created":"2026-09-30","cited_by_count":0,"area":"inference"},{"id":"2609.39601","title":"GroundingPI: A Grounding Foundation Model towards Physical Intelligence with Visual Primitives","created":"2026-09-30","cited_by_count":0,"area":null},{"id":"2609.39661","title":"The Evolution of Attention in Large Language Models: Mechanisms, Trade-offs, and Emerging Trends","created":"2026-09-30","cited_by_count":0,"area":null},{"id":"2609.39801","title":"RATIO: Reasoning Analysis and Token-level Inference Optimization for Quantized Reasoning Models","created":"2026-09-30","cited_by_count":0,"area":"lr"},{"id":"2609.39863","title":"FIGS: Evaluating Multi-Turn Sycophancy Without Penalizing Empathy","created":"2026-09-30","cited_by_count":0,"area":null},{"id":"2609.39903","title":"OSWorld-Science: A Benchmark of Computer Use Agents for Learning and Using Scientific Software","created":"2026-09-30","cited_by_count":0,"area":"mllm"},{"id":"2609.39929","title":"RoPE at the End of Its Rope? Theory, Diagnosis, and Mitigation of Long-Context Failures","created":"2026-09-30","cited_by_count":0,"area":null},{"id":"2609.40253","title":"ComputerSD: Online Self-Distillation from Real-Time Feedback for Computer-Use Agents","created":"2026-09-30","cited_by_count":0,"area":"agents"},{"id":"2609.40284","title":"cua-speedrun: Standardized Benchmarking of the Speed of Computer-Use Agents","created":"2026-09-30","cited_by_count":0,"area":"agents"},{"id":"2610.00576","title":"Gestalt: Large Multimodal Interplay Model","created":"2026-09-30","cited_by_count":0,"area":"mllm"},{"id":"2610.02123","title":"Harnessing Domain Specialists in Multimodal Mixture-of-Experts for Efficient Adaptation","created":"2026-10-01","cited_by_count":0,"area":"moe"},{"id":"2610.02568","title":"Mitigating Social Sycophancy via Pluralistic Preference Optimization","created":"2026-10-01","cited_by_count":0,"area":"alignment"}],"next_cursor":"100"}