Home
Creators
Search Recent Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

Automatically Run Prompt Chains (Patreon)

Published:
2023-09-26 16:30:54
Imported:
2026-06
Tags:
haven't archived this post yet. have a subscription? use the importer!

Content

This is a Google Colab for running a sequence of prompts.  Simply drop this into ChatGPT and ask ChatGPT to update it on your sequence.  

Here are the steps:

Make your own copy of this notebook.

Update your new copy with your OpenAI API key

Copy and paste the script into ChatGPT along with the prompt sequence you want to run and ask ChatGPT to update it accordingly.

Copy that back into the notebook and save it somewhere handy!


Files

Previews only

Google Colaboratory

Comments

Chris Barnes

You mention in the video that you can feed gpt a document to train it for helping refine the response. I would like to use this sequencing function to run a few processes for formatting, editing, and outputting a google doc. Might I be able to do this with the above?

Blazing Zebra
>>118911237

Hi Chris! First off, thanks for your support. One of my future enhancements was to be able to upload a sequence of prompts (as a .txt file) to colab and have the script run the prompt sequence that is detailed in the .txt file. This would not get at formatting it into a google doc, however. I have done a little experimenting with getting ChatGPT to generate a google doc (or output a markup that google docs can read). While I still think this is possible, it is not easily done. I don't know of anyone who has done this. If you do, let me know! Otherwise, I will keep chipping away at this :)

Chris Barnes

Roger that! What I am hopeing to do is to have a document (google doc/pdf) that has a bank of terms or formatted text I use often. Id like to include that document in the prompt to GPT in one of the steps to have it pull from that bank of text rather than include it in the prompt. OR maybe there is a way to create a library of those terms/text.

Blazing Zebra
>>118960441

Interesting. Might be a use case for a vector database, depending on how big the library of terms is.