Data Skeptic cover image

Evaluating Jokes with LLMs

Data Skeptic

00:00

How to Evaluate the Quality of a Joke

Most of the time we are using humans to evaluate creativity because we are like the ultimate judges on creativity. I start working with jokes because there is a comedian, an American comedian, his name is Joe Toppling. He works for these late night shows like Jay Leno and he's also researching. And then he started publishing papers about how to create funny jokes. His tool creates really funny jokes as humans do. But he's using humans to evaluation the results. What I thought was, okay, let's see if GPT is good enough to replace people to judge jokes. You have two things important, generation and evaluation. We can generate lots of things in a few seconds in a

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app