Posts
Papers
Trends
Maps
Mastermind
Lectures Translate
Course
Dictionary
Knowledge Management System
Papers
Language Models Can Control Their Own Attention
VBVR-Pro - A Scalable and Verifiable Suite for Native Visual Reasoning
UrbanGround - From Local Perception to Spatial Agency in a Real-Scale City
Agentic Game Development as a Verifiable Trajectory Data Engine for Scaling World Models
VoiceMem - Streaming Dual-Brain Memory for Real-Time Interaction
Annotations as Rollouts - Efficient and Scalable Reinforcement Learning for Video MLLMs
Let's Scale Step by Step - Compute-Efficient Hyperparameter Transfer for Large-Scale Mixture-of-Experts
SemaPLC - A Project-Grounded, Verification-Gated Agent Harness for PLC Code Generation
Zetta ζ - An Efficient Closed-Loop Embodied Harness for Self-Evolving Physical Intelligence
EnvHarness - Awakening Static Worlds for Agent Learning
Demystifying Agent Skills - Why They Work Until They Don't
WorldClaw - Agentic 3D Open-World Generation at Scale
N0-VTLA - Scaling Vision-Tactile-Language-Action Model with Latent Tactile Tokens
Mental World Modeling
From RLVR to RLSVR - Task Transformation Induces Self-Verifiable Rewards for Open-Ended LLM Self-Improvement
HANDBOOK.md - A Benchmark for Long-Context Agentic Instruction Following
VideoCoCo - Code-as-CoT for Physically-Consistent Video Generation via an Agentic Dual-Engine System
PhiZero - A World Model Built Around Physical Language
Qwen-UI-Agent Technical Report - Toward Next-Generation Real-World Centric Foundation GUI Agents
Metis - Memory Foundation Model
Frontis-MA1 - Training an AI4AI Model towards Recursive Self-Improvement in Machine Learning Engineering
DecoEvo - Score-Decoupled Co-Evolution of Solver and Rubric-Generator Skills in Text Space
TurboVLA - Real-Time Vision-Language-Action Model at 32 Hz on an RTX 4090 with <1 GB VRAM
HiFi-UMI - Learning Deployable Manipulation Policies from High-Fidelity UMI Data Alone
A New Role for Relevance - Guiding Corpus Interaction in Agentic Search
Rethinking Classifier-Free Guidance in On-Policy Diffusion Distillation
JarvisHub - An Open Harness for Canvas-Native Multimodal Creative Agents
ABot-World-0 - Infinite Interactive World Rollout on a Single Desktop GPU
Self Gradient Forcing - Native Long Video Extrapolation
SLAI T-Rex - Full-Parameter Post-training of the DeepSeek-V4 Family on Ascend SuperPOD
Mage-Flow - An Efficient Native-Resolution Foundation Model for Image Generation and Editing
KnowAct-GUIClaw - Know Deeply, Act Perfectly, Personal GUI Assistant with Self-Evolving Memory and Skill
LongStraw - Long-Context RL Beyond 2M Tokens under a Fixed GPU Budget
SEED - Self-Evolving On-Policy Distillation for Agentic Reinforcement Learning
Long-Horizon-Terminal-Bench - Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Grading
SearchOS-V1 - Towards Robust Open-Domain Information-Seeking Agent Collaboration
VideoChat3 - Fully Open Video MLLM for Efficient and Generalist Video Understanding
ABot-AgentOS - A General Robotic Agent OS with Lifelong Multi-modal Memory
SynthDocBench - Controlled Benchmark for Long-Context Visual Document Understanding
Jet-Long - Efficient Long-Context Extension with Dynamic Bifocal RoPE
Think Through a Bottleneck - Hourglass Reasoning for Rigorous Induction
Weak-to-Strong Generalization via Direct On-Policy Distillation
Video Generation Models are General-Purpose Vision Learners
Scalable Visual Pretraining for Language Intelligence
The Illusion of Equivalency - Statistical Characterization of Quantization Effects in LLMs
AutoMem - Automated Learning of Memory as a Cognitive Skill
Vision Pretraining for Dense Spatial Perception
OrbitQuant - Data-Agnostic Quantization for Image and Video Diffusion Transformers
Verbalizable Representations Form a Global Workspace in Language Models
AgenticSTS - A Bounded-Memory Testbed for Long-Horizon LLM Agents
Embodied.cpp - A Portable Inference Runtime of Embodied AI Models on Heterogeneous Robots
ELDR - Expert-Locality-Aware Decode Routing for PD-Disaggregated MoE Serving
MemSyco-Bench - Benchmarking Sycophancy in Agent Memory
Multi-Resolution Flow Matching - Training-Free Diffusion Acceleration via Staged Sampling
Program-as-Weights - A Programming Paradigm for Fuzzy Functions
EvoPolicyGym - Evaluating Autonomous Policy Evolution in Interactive Environments
TUA-Bench - A Benchmark for General-Purpose Terminal-Use Agents
LiveEdit - Towards Real-Time Diffusion-Based Streaming Video Editing
Scaling the Horizon, Not the Parameters - Reaching Trillion-Parameter Performance with a 35B Agent
Agentic Abstention - Do Agents Know When to Stop Instead of Act?
Wan-Streamer v0.1 - End-to-end Real-time Interactive Foundation Models
Escaping the Self-Confirmation Trap - An Execute-Distill-Verify Paradigm for Agentic Experience Learning
Improved Large Language Diffusion Models
JetSpec - Breaking the Scaling Ceiling of Speculative Decoding with Parallel Tree Drafting
In-Context World Modeling for Robotic Control
DanceOPD - On-Policy Generative Field Distillation
OPID - On-Policy Skill Distillation for Agentic Reinforcement Learning
The Verification Horizon - No Silver Bullet for Coding Agent Rewards
NatureBench - Can Coding Agents Match the Published SOTA of Nature-Family Papers
Qwen-AgentWorld - Language World Models for General Agents
EnterpriseClawBench - Benchmarking Agents from Real Workplace Sessions
Are We Ready For An Agent-Native Memory System
AOHP - An Open-Source OS-Level Agent Harness for Personalized, Efficient and Secure Interaction
Unlimited OCR Works - Welcome the Era of One-shot Long-horizon Parsing
ACE-Ego-0 - Unifying Egocentric Human and Robotic Data for VLA Pretraining
PlanBench-XL - Evaluating Long-Horizon Planning of LLM Tool-Use Agents in Large-Scale Tool Ecosystems
GateMem - Benchmarking Memory Governance in Multi-Principal Shared-Memory Agents
PerceptionDLM - Parallel Region Perception with Multimodal Diffusion Language Models
Zone of Proximal Policy Optimization - Teacher in Prompts, Not Gradients
Learning from the Self-future - On-policy Self-distillation for dLLMs
DragMesh-2 - Physically Plausible Dexterous Hand-Object Interaction with Articulated Objects
Looped World Models
LoopCoder-v2 - Only Loop Once for Efficient Test-Time Computation Scaling
Moebius - 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance
Beyond the Current Observation - Evaluating Multimodal Large Language Models in Controllable Non-Markov Games
VibeThinker-3B - Exploring the Frontier of Verifiable Reasoning in Small Language Models
GameCraft-Bench - Can Agents Build Playable Games End-to-End in a Real Game Engine?
HyGRAG - A Unified Framework for Context-Aware and Relation-Aware Graph Retrieval-Augmented Generation
FastContext - Training Efficient Repository Explorer for Coding Agents
SpatialClaw - Rethinking Action Interface for Agentic Spatial Reasoning
Memento - Reconstruct to Remember for Consistent Long Video Generation
From AGI to ASI
DreamX-World 1.0 - A General-Purpose Interactive World Model
HYDRA-X - Native Unified Multimodal Models with Holistic Visual Tokenizers
On the Scaling of PEFT - Towards Million Personal Models of Trillion Parameters
ABot-Earth 0.5 - Generative 3D Earth Model
Agents' Last Exam
Orchestra-o1 - Omnimodal Agent Orchestration
OmniDirector - General Multi-Shot Camera Cloning without Cross-Paired Data
EvoArena - Tracking Memory Evolution for Robust LLM Agents in Dynamic Environments
Crafter - A Multi-Agent Harness for Editable Scientific Figure Generation from Diverse Inputs
MiniMax Sparse Attention
Gated Attention for Large Language Models - Non-linearity, Sparsity, and Attention-Sink-Free
Rethinking the Divergence Regularization in LLM RL
Optical Reasoning - Rethinking Images as an Expressive Reasoning Medium Beyond Text
RAG-Anything - All-in-One RAG Framework
MinerU2.5 - Document Parsing Vision-Language Model
Artificial Hivemind - The Open-Ended Homogeneity of Language Models (and Beyond)
Do Transformers Need Three Projections - Systematic Study of QKV Variants
NVIDIA Nemotron 3 Ultra
SpatialWorld - Benchmarking Interactive Spatial Reasoning of Multimodal Agents in Real-World Tasks
When AI Builds Itself - Anthropic의 재귀적 자기개선 경고
Generative Criticality in Large Language Model Temperature Scaling
Agent Memory - Characterization and System Implications of Stateful Long-Horizon Workloads
Recursive Language Models
Autoregressive Diffusion World Models for Off-Policy Evaluation of LLM Agents
The Self-Correction Illusion - LLMs Correct Others but Not Themselves
ReasoningFlow - Discourse Structures for Understanding LLM Reasoning Traces
Holo3.1 - Fast and Local Computer Use Agents
MARDoc - A Memory-Aware Refinement Agent Framework for Multimodal Long Document QA
The Deterministic Horizon - When Extended Reasoning Fails and Tool Delegation Becomes Necessary
Audio Interaction Model
Quantized Reasoning Models Think They Need to Think Longer, but They Do Not
Cosmos 3 - Omnimodal World Models for Physical AI
MMG2Skill - Can Agents Distill In-the-Wild Guides into Self-Evolving Skills
HLL - Can Agents Cross Humanity's Last Line of Verification
MetaForge - A Self-Evolving Multimodal Agent that Retrieves, Adapts, and Forges Tools On Demand
awesome-architecture - 코드 대신 아키텍처를 가르치는 오픈소스 지식 베이스
Video2GUI - Synthesizing Large-Scale Interaction Trajectories for Generalized GUI Agent Pretraining
Self-Manager - Parallel Agent Loop for Long-form Deep Research
Enhancing Train-Free Infinite-Frame Generation for Consistent Long Videos
The Era of Agentic Organization - Learning to Organize with Language Models
World Action Models - The Next Frontier in Embodied AI
FlowCompile - An Optimizing Compiler for Structured LLM Workflows
Model-Agnostic Lifelong LLM Safety via Externalized Attack-Defense Co-Evolution
ProgramBench - Can Language Models Rebuild Programs From Scratch
LongMemEval-V2 - Evaluating Long-Term Agent Memory Toward Experienced Colleagues
OmniNFT - Modality-wise Omni Diffusion Reinforcement for Joint Audio-Video Generation
Nested Learning - The Illusion of Deep Learning Architectures
SenseNova-U1 - Unifying Multimodal Understanding and Generation with NEO-unify Architecture
Predictive Maps of Multi-Agent Reasoning - A Successor-Representation Spectrum for LLM Communication Topologies
Soohak - A Mathematician-Curated Benchmark for Evaluating Research-level Math Capabilities of LLMs
Subliminal Learning - Language Models Transmit Behavioral Traits via Hidden Signals in Data
HeavySkill - Heavy Thinking as the Inner Skill in Agentic Harness
KisMATH - Do LLMs Have Knowledge of Implicit Structures in Mathematical Reasoning
NVIDIA Nemotron-Personas-Korea
Recursive Multi-Agent Systems
World-R1 - Reinforcing 3D Constraints for Text-to-Video Generation
Dive into Claude Code The Design Space of AI Agent Systems
Think in Strokes, Not Pixels - Process-Driven Image Generation via Interleaved Reasoning
Representation Alignment for Just Image Transformers is not Easier than You Think
MA-EgoQA - Question Answering over Egocentric Videos from Multiple Embodied Agents
OpenClaw-RL - Train Any Agent Simply by Talking
Thinking to Recall - How Reasoning Unlocks Parametric Knowledge in LLMs
SkillOrchestra - Learning to Route Agents via Skill Transfer
OmniLottie - Generating Vector Animations via Parameterized Lottie Tokens
Kimi k2.5 - 200만 토큰의 멀티모달 에이전트
Deep Delta Learning
Seedream 4.0 Toward Next-generation Multimodal Image Generation
Souper-Model How Simple Arithmetic Unlocks State-of-the-Art LLM Performance
Black-Box On-Policy Distillation of Large Language Models
Depth Anything 3 Recovering the Visual Space from Any Views
LeJEPA Provable and Scalable Self-Supervised Learning Without the Heuristics
Kosmos An AI Scientist for Autonomous Discovery
Emu3.5 Native Multimodal Models are World Learners
Context Engineering 2.0 - The Context of Context Engineering
Kimi Linear An Expressive, Efficient Attention Architecture
Exploring Conditions for Diffusion Models in Robotic Control
Real Deep Research for AI, Robotics and Beyond
A Survey of Data Agents Emerging Paradigm or Overstated Hype
A Definition of AGI
The Free Transformer
FineVision Open Data Is All You Need
DeepSeek-OCR Contexts Optical Compression
Detect Anything via Next Point Prediction
The Dragon Hatchling The Missing Link between the Transformer and Models of the Brain
MCPMark A Benchmark for Stress-Testing Realistic and Comprehensive MCP Use
VFF-Net Evolving forward–forward algorithms into convolutional neural networks for enhanced computational insights
Diffusion Transformers with Representation Autoencoders
Training-Free Group Relative Policy Optimization
Pushing on Multilingual Reasoning Models with Language-Mixed Chain-of-Thought
Meta-Awareness Enhances Reasoning Models Self-Alignment Reinforcement Learning
Agent Learning via Early Experience
FAST-DLLM V2 Efficient Block-Diffusion LLM
Less is More Recursive Reasoning with Tiny Networks
Efficient Multi-modal Large Language Models via Progressive Consistency Distillation
CoDA Agentic Systems for Collaborative Data Visualization
Video models are zero-shot learners and reasoners
Soft Tokens, Hard Truths
Sharing is Caring Efficient LM Post-Training with Collective RL Experience Sharing
Why Language Models Hallucinate
Hunyuan3D Studio End-to-End AI Pipeline for Game-Ready 3D Asset Generation
DINOv3
Prefix-Tuning Optimizing Continuous Prompts for Generation
You Only Look Once, Unified Real-Time Object Detection
Attention Is All You Need
EXAONE 4.0 Unified Large Language Models Integrating Non-reasoning and Reasoning Modes
군중 상황에서 정확한 다중 사람의 자세 인식을 위한 군중 자세 주석 데이터 세트
