| Linearizing Vision Transformer with Test-Time Training | Image & video gen | Yining Li +5 | May 28, 2026 | 20 |
| Is Position Bias in Dense Retrievers Built In-or Learned from Data? | Safety & alignment | Daegon Yu +2 | May 26, 2026 | 20 |
| How and What to Imagine? Visual Thinking in Unified Multimodal Models for Cross-View Spatial Reasoning | Vision & multimodal | Qian Yang +6 | May 26, 2026 | 20 |
| Negligible in Size, Significant in Effect: On Scale Vectors in Large Language Models | Safety & alignment | Mingze Wang +5 | May 26, 2026 | 20 |
| Balancing Fidelity and Diversity in Diffusion Models via Symmetric Attention Decomposition: Hopfield Perspective | Image & video gen | Hyunmin Cho +2 | May 26, 2026 | 20 |
| Enhancing In-context Panoramic Generation via Geometric-aware Pretraining | Efficiency & systems | Haoran Feng +4 | Jul 9, 2026 | 19 |
| AutoMem: Automated Learning of Memory as a Cognitive Skill | Agents | Shengguang Wu +4 | Jul 1, 2026 | 19 |
| CausalMix: Data Mixture as Causal Inference for Language Model Training | Vision & multimodal | Zinan Tang +8 | Jul 1, 2026 | 19 |
| ABot-M0.5: Unified Mobility-and-Manipulation World Action Model | Robotics | Ronghan Chen +20 | Jul 1, 2026 | 19 |
| Xiaomi-GUI-0 Technical Report | Vision & multimodal | Wanxia Cao +29 | Jun 30, 2026 | 19 |
| IV-CoT: Implicit Visual Chain-of-Thought for Structure-Aware Text-to-Image Generation | Image & video gen | Zixuan Li +12 | Jun 23, 2026 | 19 |
| Text-Vision Co-Instructed Image Editing | Robotics | Chenxi Xie +3 | Jun 15, 2026 | 19 |
| Reroute, Don't Remove: Recoverable Visual Token Routing for Vision-Language Models | Vision & multimodal | Cheng-Yu Yang +2 | Jun 10, 2026 | 19 |
| ComBench: A Benchmark for Rigorous Proof Reasoning and Constructive Realization in Olympiad-Level Combinatorics | Data & benchmarks | Shunkai Zhang +17 | Jun 9, 2026 | 19 |
| OmniGameArena: A Unified UE5 Benchmark for VLM Game Agents with Improvement Dynamics | Vision & multimodal | Mingxian Lin +11 | Jun 8, 2026 | 19 |
| SubtleMemory: A Benchmark for Fine-Grained Relational Memory Discrimination in Long-Horizon AI Agents | Agents | Wenxuan Wang +6 | Jun 4, 2026 | 19 |
| MLEvolve: A Self-Evolving Framework for Automated Machine Learning Algorithm Discovery | Agents | Shangheng Du +13 | Jun 4, 2026 | 19 |
| Personal AI Agent for Camera Roll VQA | Vision & multimodal | Thao Nguyen +4 | Jun 3, 2026 | 19 |
| OpenWebRL: Demystifying Online Multi-turn Reinforcement Learning for Visual Web Agents | Vision & multimodal | Rui Yang +9 | Jun 1, 2026 | 19 |
| LongLive-RAG: A General Retrieval-Augmented Framework for Long Video Generation | Image & video gen | Qixin Hu +4 | Jun 1, 2026 | 19 |
| ESPO: Early-Stopping Proximal Policy Optimization | Reinforcement learning | Zihang Li +10 | May 28, 2026 | 19 |
| Advancing Creative Physical Intelligence in Large Multimodal Models | Vision & multimodal | Cheng Qian +12 | May 25, 2026 | 19 |
| VitaBench 2.0: Evaluating Personalized and Proactive Agents in Long-Term User Interactions | Agents | Yuxin Chen +13 | May 26, 2026 | 19 |
| Recursive Flow Matching | Image & video gen | Jiahe Huang +3 | May 26, 2026 | 19 |
| Long-Horizon-Terminal-Bench: Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Grading | Vision & multimodal | Zongxia Li +12 | Jul 9, 2026 | 18 |
| From Foundation to Application: Improving VLA Models in Practice | Robotics | Wei Wu +23 | Jul 7, 2026 | 18 |
| Look Before You Leap: Distilling Tree Search into Action Evaluation for Frozen VLA Models | Robotics | Xinyi Xie +9 | Jul 4, 2026 | 18 |
| SkillCoach: Self-Evolving Rubrics for Evaluating and Enhancing Agentic Skill-Use | Agents | Jiayin Zhu +6 | Jul 2, 2026 | 18 |
| PACE: A Proxy for Agentic Capability Evaluation | Agents | Yueqi Song +10 | Jul 2, 2026 | 18 |
| Logit-Contribution Scoring Identifies Non-Literal Retrieval Heads | Vision & multimodal | Aryo Pradipta Gema +2 | Jul 1, 2026 | 18 |
| Perceive-to-Reason: Decoupling Perception and Reasoning for Fine-Grained Visual Reasoning | Vision & multimodal | Hongxing Li +13 | Jul 1, 2026 | 18 |
| Running the Gauntlet: Re-evaluating the Capabilities of Agents Beyond Familiar Environments | Vision & multimodal | Mykola Vysotskyi +24 | Jun 25, 2026 | 18 |
| Why Multi-Step Tool-Use Reinforcement Learning Collapses and How Supervisory Signals Fix It | Agents | Yupu Hao +4 | Jun 24, 2026 | 18 |
| Autodata: An agentic data scientist to create high quality synthetic data | Agents | Ilia Kulikov +14 | Jun 24, 2026 | 18 |
| The Hitchhiker's Guide to Agentic AI: From Foundations to Systems | Agents | Haggai Roitman | Jun 22, 2026 | 18 |
| Self-Compacting Language Model Agents | Agents | Tianjian Li +7 | Jun 22, 2026 | 18 |
| HydraHead: From Head-Level Functional Heterogeneity to Specialized Attention Hybridization | Safety & alignment | Zhentao Tan +6 | Jun 18, 2026 | 18 |
| Learning from Your Own Mistakes: Constructing Learnable Micro-Reflective Trajectories for Self-Distillation | Reinforcement learning | Zhilin Huang +8 | Jun 17, 2026 | 18 |
| GateMem: Benchmarking Memory Governance in Multi-Principal Shared-Memory Agents | Agents | Zhe Ren +9 | Jun 17, 2026 | 18 |
| SproutRAG: Attention-Guided Tree Search with Progressive Embeddings for Long-Document RAG | Code | Amirhossein Abaskohi +3 | Jun 16, 2026 | 18 |
| BadWorld: Adversarial Attacks on World Models | Image & video gen | Linghui Shen +2 | Jun 15, 2026 | 18 |
| Grammar-Constrained Decoding Can Jailbreak LLMs into Generating Malicious Code | Code | Yitong Zhang +2 | Jun 10, 2026 | 18 |
| TRACE: A Unified Rollout Budget Allocation Framework for Efficient Agentic Reinforcement Learning | Agents | Heming Zou +11 | Jun 9, 2026 | 18 |
| EEVEE: Towards Test-time Prompt Learning in the Real World for Self-Improving Agents | Agents | Weixian Xu +2 | Jun 9, 2026 | 18 |
| The Road Ahead in Autonomous Driving: The KITScenes Multimodal Dataset | Robotics | Richard Schwarzkopf +23 | Jun 1, 2026 | 18 |
| Domain-Specific Data Synthesis for LLMs via Minimal Sufficient Representation Learning | Code | Tong Ye +8 | May 28, 2026 | 18 |
| From Prompt Injection to Persistent Control: Defending Agentic Harness Against Trojan Backdoors | Speech & audio | Jiejun Tan +6 | May 29, 2026 | 18 |
| CausaLab: A Scalable Environment for Interactive Causal Discovery Toward AI Scientists | Robotics | Junlin Yang +9 | May 25, 2026 | 18 |
| DrugGen 2: A disease-aware language model for enhancing drug discovery | Reinforcement learning | Ali Motahharynia +6 | Jul 9, 2026 | 17 |
| TurnOPD: Making On-Policy Distillation Turn-Aware for Efficient Long-Horizon Agent Training | Agents | Yuhang Zhou +5 | Jul 7, 2026 | 17 |
| Unified Audio Intelligence Without Regressing on Text Intelligence | Speech & audio | Zhifeng Kong +19 | Jul 6, 2026 | 17 |
| Discretizing Reward Models | Reinforcement learning | Vijay Viswanathan +6 | Jun 19, 2026 | 17 |
| Native Active Perception as Reasoning for Omni-Modal Understanding | Speech & audio | Zhenghao Xing +10 | Jun 17, 2026 | 17 |
| SAE Interventions are Unreliable: Post-Intervention Recovery of Suppressed Behavior | Safety & alignment | Mingyue Cui +2 | Jun 16, 2026 | 17 |
| Rethinking the Role of Efficient Attention in Hybrid Architectures | LLMs & reasoning | Ziqing Qiao +8 | Jun 13, 2026 | 17 |
| OneRank: Unified Transformer-Native Ranking Architecture for Multi-Task Recommendation | Code | Jiakai Tang +10 | Jun 15, 2026 | 17 |
| Memento: Reconstruct to Remember for Consistent Long Video Generation | Image & video gen | Xuan Wei +7 | Jun 12, 2026 | 17 |
| RepFusion: Leveraging Multimodal Priors for Denoising in Representation Space | Image & video gen | Xichen Pan +5 | Jun 12, 2026 | 17 |
| Complexity-Balanced Diffusion Splitting | Image & video gen | Noam Issachar +2 | Jun 4, 2026 | 17 |
| MapAgent: An Industrial-Grade Agentic Framework for City-scale Lane-level Map Generation | Vision & multimodal | Deguo Xia +8 | Jun 3, 2026 | 17 |
| KletterMix: Climbing Toward High-Quality German Pretraining Data | Efficiency & systems | Maurice Kraus +6 | Jun 2, 2026 | 17 |
| MemTrain: Self-Supervised Context Memory Training | Agents | Ziheng Li +4 | Jun 2, 2026 | 17 |
| TRON: Targeted Rule-Verifiable Online Environments for Visual Reasoning RL | Vision & multimodal | Tianze Yang +5 | Jun 1, 2026 | 17 |
| When Does Multi-Agent RL Improve LLM Workflows? Workflow, Scale, and Policy-Sharing Tradeoffs | Agents | Yifan Zeng +6 | May 22, 2026 | 17 |
| LiteCoder-Terminal: Scaling Long-Horizon Terminal Environments for Learning Language Agents | Agents | Xiaoxuan Peng +7 | May 28, 2026 | 17 |
| WorldMemArena: Evaluating Multimodal Agent Memory Through Action-World Interaction | Vision & multimodal | Chengzhi Liu +16 | May 28, 2026 | 17 |
| LiveBrowseComp: Are Search Agents Searching, or Just Verifying What They Already Know? | Agents | HuiMing Fan +7 | May 27, 2026 | 17 |
| GE-Sim 2.0: A Roadmap Towards Comprehensive Closed-loop Video World Simulators for Robotic Manipulation | Robotics | Boxiang Qiu +14 | May 26, 2026 | 17 |
| Fast-dDrive: Efficient Block-Diffusion VLM for Autonomous Driving | Robotics | Kewei Zhang +11 | May 22, 2026 | 17 |
| EdgeBench: Unveiling Scaling Laws of Learning from Real-World Environments | Agents | Deyao Zhu +39 | Jul 6, 2026 | 16 |
| dOPSD: On-Policy Self-Distillation for Diffusion Language Models | Image & video gen | Phuong Tuan Dat +2 | Jul 5, 2026 | 16 |
| Perceptual Flow Matching for Few-Step Generative Modeling | Image & video gen | Chuyang Zhao +9 | Jul 3, 2026 | 16 |
| Measuring the Gap Between Human and LLM Research Ideas | Data & benchmarks | Ziyu Chen +2 | Jul 1, 2026 | 16 |
| Optimizing Visual Generative Models via Distribution-wise Rewards | Vision & multimodal | Ruihang Li +6 | Jul 2, 2026 | 16 |
| AGVBench: A Reliability-Oriented Benchmark of Data Augmentation for Vein Recognition | Vision & multimodal | Haiyang Li +6 | Jul 2, 2026 | 16 |
| RedVox: Safety and Fairness Gaps in Speech Models Across Languages | Speech & audio | Beatrice Savoldi +4 | Jun 25, 2026 | 16 |
| Play2Perfect: What Matters in Dexterous Play Pretraining for Precise Assembly? | Robotics | Tyler Ga Wei Lum +3 | Jun 24, 2026 | 16 |
| GUICrafter: Weakly-Supervised GUI Agent Leveraging Massive Unannotated Screenshots | Vision & multimodal | Sunqi Fan +6 | Jun 29, 2026 | 16 |
| DreamForge-World 0.1 Preview: A Low-Compute Real-Time Controllable World Model | Image & video gen | Daniyel Ayupov +1 | Jun 29, 2026 | 16 |
| EBench: Elemental Diagnosis of Generalist Mobile Manipulation Policies | Robotics | Ning Gao +24 | Jun 20, 2026 | 16 |
| Current World Models Lack a Persistent State Core | Image & video gen | Jinpeng Lu +10 | Jun 18, 2026 | 16 |
| Context-Aware RL for Agentic and Multimodal LLMs | Vision & multimodal | Peiyang Xu +6 | Jun 15, 2026 | 16 |
| Beyond Monolingual Deep Research: Evaluating Agents and Retrievers with Cross-Lingual BrowseComp-Plus | Agents | Yuheng Lu +8 | Jun 13, 2026 | 16 |
| TokenPilot: Cache-Efficient Context Management for LLM Agents | Agents | Buqiang Xu +14 | Jun 15, 2026 | 16 |
| Retrieve, Don't Retrain: Extending Vision Language Action Models to New Tasks at Test Time | Robotics | Jeongeun Park +5 | Jun 14, 2026 | 16 |
| MVEB: Massive Video Embedding Benchmark | Speech & audio | Adnan El Assadi +15 | Jun 12, 2026 | 16 |
| Pythagoras-Prover: Advancing Efficient Formal Proving via Augmented Lean Formalisation | Image & video gen | Joshua Ong Jun Leang +7 | Jun 10, 2026 | 16 |
| DyCo-RL: Dynamic Cross-Modal Coordination for Visual Reasoning | Vision & multimodal | Hangui Lin +8 | Jun 6, 2026 | 16 |
| One Token per Multimodal Evidence: Latent Memory for Resource-Constrained QA | Vision & multimodal | Zhi Zheng +4 | Jun 9, 2026 | 16 |
| LIMMT: Less is More for Motion Tracking | Science & bio | Yu Guan +9 | Jun 5, 2026 | 16 |
| LLM Explainability with Counterfactual Chains and Causal Graphs | Safety & alignment | Nirit Nussbaum-Hoffer +3 | Jun 4, 2026 | 16 |
| Dream.exe: Can Video Generation Models Dream Executable Robot Manipulation? | Robotics | Rui Zhao +8 | Jun 4, 2026 | 16 |
| Eliciting Complex Spatial Reasoning in MLLMs through Wide-Baseline Matching | Vision & multimodal | Hao Zhong +10 | Jun 2, 2026 | 16 |
| Filter, Then Reweight: Rethinking Optimization Granularity in On-Policy Distillation | Code | Yuying Li +8 | Jun 1, 2026 | 16 |
| MCP-Persona: Benchmarking LLM Agents on Real-World Personal Applications via Environment Simulation | Agents | Wenhao Wang +11 | Jun 1, 2026 | 16 |
| Skill is Not One-Size-Fits-All: Model-Aware Skill Alignment for LLM Agents | Agents | Jianxiang Yu +5 | May 29, 2026 | 16 |
| StreamChar: Long-Horizon Streaming Character Audio-Video Generation with Decoupled Orchestration | Speech & audio | Linrui Tian +2 | May 25, 2026 | 16 |
| Towards Verifiable Multimodal Deep Research: A Multi-Agent Harness for Interleaved Report Generation | Vision & multimodal | Chenghao Zhang +4 | May 28, 2026 | 16 |
| AsyncTool: Evaluating the Asynchronous Function Calling Capability under Multi-Task Scenarios | Agents | Kou Shi +9 | May 27, 2026 | 16 |
| PRISM: A Multi-Dimensional Benchmark for Evaluating LLM Peer Reviewers | Data & benchmarks | Ngoc Phan Phuoc Loc +10 | May 26, 2026 | 16 |