AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Do Language Models Understand?
GPT-3 can do addition and multiplication, but not 100% of the time. What differentiates the one it gets correct in? Do things that it doesn't get correct in. How should we calibrate what things these should be deployed for and what things they should not be deployed for based on the institutions? Those are the kind of things I'm really interested in.