https://jamesakl.com/posts/cuda-ontology/

From Solving to Verifying: A Unified Objective for Robust Reasoning in LLMs https://arxiv.org/abs/2511.15137 https://www.alphaxiv.org/ru/overview/2511.15137v1

Olmo 3 https://www.datocms-assets.com/64837/1763646865-olmo3technicalreport-1.pdf

Seer: Online Context Learning for Fast Synchronous LLM Reinforcement Learning https://arxiv.org/abs/2511.14617 https://www.alphaxiv.org/ru/overview/2511.14617v1

https://pytorch.org/blog/pina-joins-the-pytorch-ecosystem-a-unified-framework-for-scientific-machine-learning/

Мне кажется они умеют понижать IQ агентским системам