The papers span large model evaluation complex process reasoning reinforcement learning competition-level mathematics and generative recommendation systems at the premier NLP conference.