This Prompt Engineering Secret Is Too Good To Give Away

This LLM prompting secret allows you to make chain of thought prompts with generations from one response from an AI being fed to another prompt, which is called chain of thought prompting.

However, the interesting thing about this specifically is that we can achieve chain of thought prompting using XML tags to denote which part of the prompt refers to which large output of information

One of the biggest issues with LLMs is prompt engineering is that the prompts become so large, with such massive amounts of information in them, that the LLM gets confused - this is where XML Prompt engineering comes into play - instead of worrying about whether the LLM can understand your complicated promptss, you simply wrap large amounts of data or text in XML tags, then refer to the XML tags throughout your prompting process

This makes prompt engineering and chain of thought prompting significantly easier, and allows you to easily feed the output of one prompt into the next prompt, allowing you to create reasoning without the use of o1-preview or o1-pro

Inside large SaaS projects or large codebases like Harbor, this prompt engineering technique is an essential tool to creating workflows that allow a high degree of accuracy

Thanks for watching and PEACE

Hamish

Try our SEO tool: https://harborseo.ai/
Work with us: https://calendly.com/incomestreamsurfers-strategy-session/seo

prompt engineering,chatgpt prompts,claude prompts,chatgpt 4o mini,chatgpt 4o,chatgpt pro,prompt engineering tips,prompt engineering 101,ai