AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Is GPT-3 Unable to Deal With Semantics?
GPT-3 is able to do some basic math and two digit numbers. It basically fails if you take larger digit numbers, so four digit numbers and so on makes carrying mistakes. If you take larger numbers, you don't get useful results at all. And this could be an issue of the training set. There are not many examples of successful long form addition and standard human written text.