TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Everything I'll forget about prompting LLMs

5 pointsby madmax108over 1 year ago

1 comment

spdustinover 1 year ago
This is very well done. There’s success to be had in asking the model to list its own steps in the preamble of the output, too. “Give it space to think”. In the context of an app wrapped around the LLM, you can ask it to wrap those steps in a <plan> tag, and then trim those plans out of the conversation context (and display!) when the completion has returned.