How (And Why) To Save Your ChatGPT/LLM Outputs / Generations / Completions
For all the (deserved) attention that prompt engineering gets, remarkably little thought gets given to how to save the things that come out of large language models, like ChatGPT. Today, I share a couple of tips and tricks for saving and organising LLM outputs .. and cover why you might want to.
Timestamps:
Intro - 00:00
The neglected state of LLM output storage - 01:20
LLM outputs are often in markdown - 02:40
Outputting to JSON - 03:10
Nuclino note editor - 03:55
Whisper AI (Chrome extension) - 05:00
Voice to text dictation tools for prompting - 06:00
Using saved outputs for new prompts - 08:30
Using an LLM to review previous generations - 09:00
Structuring saved prompts and outputs - 11:25
Reworking old outputs - 13:00
Saving outputs into Google Drive - 15:20
Enabling markdown in Google Drive - 15:50
#promptengineering #largelanguagemodels