
16 - Preparing for Debate AI with Geoffrey Irving
AXRP - the AI X-risk Research Podcast
How Can That Be Right?
The goal is to improve our ability to collect accurate data from humans. Andf you get this right, sotif two things happens. The one is, if you have a give a fixed budget of human data, you can collect more of it effectively. You then get to more accurate classifiers, and therefore catch more safety failures,. do a better job. If i use the ward model to decline to answer some of the time, that to 80 % out of the times that you answer. That's right. At other times we answer, it is if you skip a third of the questions to its tay, it is relatively useful, but it is now significantly more accurate. But
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.