Deeper is Not Always Better: Mitigating the Alignment Tax via Confident Layer Decoding Paper • 2606.21906 • Published 12 days ago • 24
VibeThinker-3B: Exploring the Frontier of Verifiable Reasoning in Small Language Models Paper • 2606.16140 • Published 17 days ago • 121
view article Article LeRobot Humanoid: An Open, Low-Cost, 3D-Printed Humanoid for Robot Learning VirgileBatto • May 21 • 63