All Posts
/
Oct 6, 2026
Looped DiT, Looped Models Done Right, Looped For Free, and Triadic Linear Attention
Sep 29, 2026
Plus more about Memory Attention, JEPA-Anything, and Contrastive World Models
Sep 24, 2026
Plus more about Score Centering, Learning From Hard Problems, and FlashREINFORCE
Sep 15, 2026
Plus more about Thinking with Looped Flows, Recurrent Looped Transformers, and Don't Do Dropouts
Sep 9, 2026
Plus more about TTPO, Normalized LoRA, and Fast Weight Attention for Continual Learning
Sep 1, 2026
Plus more about LeVJEPA, Code World Model, and Gemini's AI Scientist
Aug 25, 2026
plus more about SPADE, skill issue, and Q-learning with World Models
Aug 19, 2026
Plus more about Full-Bandwidth Transformer, Small Scaling Law, and SFT Conflicts RL Coexists
Aug 11, 2026
plus more about Leanstral, Towards Physics of Multimodal Pretraining, and loss does not see the basis
