**Zero-Shot Chain-of-Thought (COT)** is a [[Prompt Engineering|prompting]] strategy that encourages [[Large Language Model|language models]] to break down complex problems by explicitly asking them to "think step by step" before providing a final answer. This approach leverages the model's inherent reasoning capabilities without providing any examples. ## Structure The basic structure follows this pattern: - Task instruction - Question - Explicit request: *"Let's think step by step"* ## Example From the [[MAPWise Evaluating Vision-Language Models for Advanced Map Queries|MapWise benchmark]]: ``` Instruction - Your task is to answer the question based on the provided Image. Question - Count the states bordering Canada with a value lower than New Mexico? Output - Let's think step by step, explain the steps and then provide the final answer. ```