Prompt Engineering Guide
by Satish Yenumula
There are >1500 papers in arXiv about prompt engineering. Every new model that comes out or if someone finds a better way to write a prompt and achieve better results, it is given a name and called a certain way of prompting. This has led to various techniques such as few-shot learning, chain of trees, self-criticism, planning strategies and all sorts of other names that are difficult to keep track of and end up being counter productive.
Lost in the world of prompting, is when I came across this paper: The Prompt Report: A Systematic Survey of Prompt Engineering Techniques
Key Techniques
Few-Shot Learning
Description of few-shot learning techniques…
Chain of Thought
How chain of thought prompting works…
Self-Criticism
Methods for self-improving prompts…
Conclusion
This guide provides a systematic approach to prompt engineering…
tags: AI - Prompting - Guide