PackForcing: Short Video Training Suffices for Long Video Sampling and Long Context Inference Paper • 2603.25730 • Published Mar 26 • 51
MIND: Benchmarking Memory Consistency and Action Control in World Models Paper • 2602.08025 • Published Feb 8 • 14
SIMART: Decomposing Monolithic Meshes into Sim-ready Articulated Assets via MLLM Paper • 2603.23386 • Published Mar 24 • 41
WildWorld: A Large-Scale Dataset for Dynamic World Modeling with Actions and Explicit State toward Generative ARPG Paper • 2603.23497 • Published Mar 24 • 93
Astrolabe: Steering Forward-Process Reinforcement Learning for Distilled Autoregressive Video Models Paper • 2603.17051 • Published Mar 17 • 69
Speed by Simplicity: A Single-Stream Architecture for Fast Audio-Video Generative Foundation Model Paper • 2603.21986 • Published Mar 23 • 125