Blog
No more endless PDFs. Discover the core value of the latest top-tier research in one article.
The Finetuner's Fallacy: When to Pretrain with Your Finetuning Data
Efficient Exploration at Scale
When AI Navigates the Fog of War
Transformers are Bayesian Networks
MosaicMem: Hybrid Spatial Memory for Controllable Video World Models
Fast-WAM: Do World Action Models Need Test-time Future Imagination?
GigaWorld-Policy: An Efficient Action-Centered World--Action Model
Delightful Policy Gradient
Understanding Reasoning in LLMs through Strategic Information Allocation under Uncertainty
OpenSeeker: Democratizing Frontier Search Agents by Fully Open-Sourcing Training Data
PRISM: Demystifying Retention and Interaction in Mid-Training
MDM-Prime-v2: Binary Encoding and Index Shuffling Enable Compute-optimal Scaling of Diffusion Language Models
WorldCam: Interactive Autoregressive 3D Gaming Worlds with Camera Pose as a Unifying Geometric Representation
Grounding World Simulation Models in a Real-World Metropolis
FASTER: Rethinking Real-Time Flow VLAs
Unified Spatio-Temporal Token Scoring for Efficient Video VLMs
Cubic Discrete Diffusion: Discrete Visual Generation on High-Dimensional Representation Tokens
Code-A1: Adversarial Evolving of Code LLM and Test LLM via Reinforcement Learning
CoVerRL: Breaking the Consensus Trap in Label-Free Reasoning via Generator-Verifier Co-Evolution
M$^2$RNN: Non-Linear RNNs with Matrix-Valued States for Scalable Language Modeling
World Reconstruction From Inconsistent Views
Astrolabe: Steering Forward-Process Reinforcement Learning for Distilled Autoregressive Video Models
Stereo World Model: Camera-Guided Stereo Video Generation
Adaptive Auxiliary Prompt Blending for Target-Faithful Diffusion Generation
Look Before Acting: Enhancing Vision Foundation Representations for Vision-Language-Action Models
Loc3R-VLM: Language-based Localization and 3D Reasoning with Vision-Language Models
AI Scientist via Synthetic Task Scaling
How do LLMs Compute Verbal Confidence
HorizonMath: Measuring AI Progress Toward Mathematical Discovery with Automatic Verification
Eigenforms and graphs of Hecke operators with wild ramification