(All images are by the author unless otherwise noted)
Introduction
Quick engineering is the practice of designing and refining prompts (text inputs) to improve the behavior of large language models (LLMs). The goal is to obtain the desired responses from the model by carefully crafting the instructions. The most used incitement techniques are:
- Chain of thought: It involves generating a step-by-step reasoning process to reach a conclusion. The model is pushed to “think out loud” by explicitly stating the logical steps that lead to the final answer.
- React (Reason+Act): Combine reasoning with action. The model not only thinks about a problem but also takes actions based on its reasoning. Therefore, it is more interactive as the model alternates between reasoning steps and actions, refining its approach iteratively. Basically, it is a loop of “thinking”, “action”, “observing”.
Let's do an example: imagine asking an ai to “find the best laptop under $1000.”
– Normal response: “Lenovo Thinkpad”.
– Chain of thought response: “I need to consider factors like performance, battery life, and…