Chain of Thought

Wei et al. (2022) showed that appending one phrase to a prompt dramatically improves LLM reasoning. This phrase became iconic. Note: newer models like OpenAI's o1/o3 and Claude's extended thinking have CoT built-in. This challenge is about the original prompting technique. What is the exact phrase? Flag format: CONGRESS{[the exact phrase, lowercase]} Example: CONGRESS{please explain your reasoning}