Large Language Models Are Biased and Inaccurate. Can They Be Improved?

12:00pm - 12:10pm on Saturday, October 7

Bard Fetler

Watch:

https://youtu.be/8FXb47du0hI

Description

ChatGPT, Google Bard, and other large language models are all the rage now, and attracting a lot of attention. But they are known to not always produce truthful or accurate text. Why is that?

We’ll explore some ideas behind the models and data, and possible ways to improve model behavior.