DOTS: Learning to Reason Dynamically in LLMs via Optimal Reasoning Trajectories Search Paper • 2410.03864 • Published 12 days ago • 9