[sergiotapia]

I also rather like Claude's capability of using XML tags as scrap-pads of sorts to see the inner workings of the AI. Makes for better prompt design.

https://docs.anthropic.com/claude/docs/give-claude-room-to-t...

You can see how the AI arrived at the response.

[leodriesch]

This is just a prompting hack you can use with any LLM, not exclusive to Claude. But I do like the fact that they include these tricks in their documentation.