LLM Hackathon: Evals and Red Teaming

Located at Meeting Room 1.6.

For this month’s event, we’re going to be evaluating or red-teaming a language model! Before the event, have a think about something you want to test a language model like ChatGPT on. At the event, you can get support from other participants to develop this idea into a quick prototype.

While being able to use the API is handy for some evaluations, you don’t need programming experience—if you want to just use the web app for this, you can. I’m happy to help people get set up with the OpenAI API though!

If you want to use the API, set up an account ahead of time just in case. I’ll also have a GPT-4 subscription for those who want to test stuff on GPT-4.