We investigate the reasons masking diffusion dominates discrete diffusion.

Emergence of an alignment between LLMs' and the brain's computational dynamics, and key factors allowing it : scaling and context size.

We introduce a new class of Reinforcement Learned Teachers trained to provide effective reasoning traces for downstream distillation, yielding more effective data for distillation and cold-starting than orders of magnitude larger reasoning LMs.

We present COS3D, a new collaborative prompt-segmentation framework that effectively contributes to integrating complementary language and segmentation cues throughout the entire pipeline.

We introduce ALE-bench, a new benchmark for evaluating AI systems on score-based algorithmic programming contests.

We train RL agents directly from high-level specifications, without reward functions or domain-specific oracles.