3 Comments
Jan 20, 2023Liked by Andrew Mao

Hey Andrew, this is a crisp description of your LLM use cases at Parsnip and the tech seems really impactful. I’m curious how you are managing content moderation (given there is a lot of food related content that can be triggering in the general corpus) and if you will set policies for LLM use. Might be interesting for future posts

Dianne Na Penn

Expand full comment
author

Hey Dianne, what's an example of triggering food-related content that you have in mind?

Expand full comment

Hi Andrew, I am thinking words which may be triggering to certain user groups like those facing weight or health issues. You may already be thinking about this but words like "big appetite" or "generous portion" have some connotations here. ChatGPT responses tend to utilize general corpus and for general text use cases may not be well tuned and apply a moderation policy on this area.

Expand full comment