VersatileFFN: Achieving Parameter Efficiency in LLMs via Adaptive Wide-and-Deep Reuse Paper • 2512.14531 • Published 10 days ago • 11
From Next-Token to Next-Block: A Principled Adaptation Path for Diffusion LLMs Paper • 2512.06776 • Published 19 days ago • 24
From Next-Token to Next-Block: A Principled Adaptation Path for Diffusion LLMs Paper • 2512.06776 • Published 19 days ago • 24