100 Days of Gen AI: Day 19

I spent the morning reading about OpenAI’s tools for Fine Tuning, Model Distillation and Evals. I think model distillation could be a way to make my planned “Terraform Cloud at Benchling” chatbot more cost efficient (not that cost efficiency is particularly critical for the volume of requests I expect it to get).

I’ve also been thinking about a future where content moderation (via LLMs) is a standard feature on our digital platforms, which (in the most hopeful case) could lead to a kinder internet.

I wonder, what kind of content moderation tools exist out there today? OpenAI actually has a post on this topic: Using GPT-4 for content moderation. This blog post was also instructive about the content moderation space and its various technical approaches. It also looks like OpenAI released a content moderation endpoint in 2022, which is documented in its latest form here.

Not a lot of coding today, so I’m feeling less accomplished, but this is still learning!