Introduction

Prompt Engineering is a discipline in formation for developing prompts to interact with large language models LLMs. As with many other disciplines related to computer science there are two trades of the art, research and development. Its having its different goals and requirements. Research goals entail enchancing capabilities of large language models and enabling new application use cases. The purpose of prompts to produce new and sometimes unknown patterns of information responses. In contrast prompts in development is all about predictability and integrating LLMs to applications, restricting output within acceptance requirements, and incorporating information from private knowledge bases.

This guide includes techniques for crafting prompts either for research or development and covers both basic and advanced to cutting edge research. It also includes a basic guide around the operation of LLMs and some historical references and their connections to RNNs and LSTMs. Therefore this guide's purpose is to be of use to both a prompt developer and a prompt researcher during their journey into this new emerging discipline of artificial intelligence.

This work is still in development and given the rapidly developing nature of the field it might be possible that not all angles are covered. If you are interested to contribute via our github repository, please send an email to prompt@promptility.ch .