Research

Chain-of-thought prompting has become a standard technique for improving LLM reasoning on complex tasks. We review the key findings, explain where the gains come from, and identify the limits the results do not always surface.

12 Apr 2025