AdaCoT: Pareto-Optimal Adaptive Chain-of-Thought Triggering via Reinforcement Learning Paper • 2505.11896 • Published 25 days ago • 57