6 events hosted · 1 upcoming
Cosmin goes over Score Centering Stabilizes Off-policy Reinforcement Learning
Yuri introduces diffusion foundations and covers "Feynman-Kac Correctors"
Rob and Cosmin go over "Full Bandwidth Transformer"
Andrey presents "Refusal geometry reflects refusal training"
Charles reviews the Anthropic J-Space Paper
Cosmin reviews the Qwen3.8-flash-next report
Q&A with Yevgen on RoboTTT: Context Scaling for Robot Policies
