Prompting for an LLM is a technique used to elicit specific information or generate desired responses from the model by providing it with a well-structured and contextually appropriate input. This input, known as a prompt, typically consists of a written or spoken text that serves as an instruction or query. LLMs like GPT-3.5 are highly versatile and capable of performing a wide range of language-related tasks, (see previous post). When done well, it allows users to harness the power of these models effectively, directing them towards achieving specific objectives by framing their requests in a clear and understandable manner.
Let’s break things down, and get into the details of how to construct prompts for an LLM.
If you really want to dig into something more comprehensive, I highly recommend this course. This free course is partly taught by one of the leader in Artificial Intelligence: Andrew Ng