SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features Paper • 2502.14786 • Published 1 day ago • 82
MoM: Linear Sequence Modeling with Mixture-of-Memories Paper • 2502.13685 • Published 3 days ago • 27
Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention Paper • 2502.11089 • Published 6 days ago • 129
Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach Paper • 2502.05171 • Published 14 days ago • 113