September 13, 2026 Machine Learning Chain-of-Thought Unlocked: Gradients are the Secret Sauce for Smarter LLM Reasoning Latest 1 papers on chain-of-thought reasoning: Sep. 13, 2026