Zero-shot prompts. The model generates responses to new prompts according to typical instruction without the need of unique examples.This “chain of assumed”, characterized with the pattern “question ? intermediate problem ? stick to-up concerns ? intermediate dilemma ? stick to-up questions ? … ? closing remedy”, guides the LLM to reach t