Natural language benchmarks don’t measure AI models’ general knowledge well

Re:Natural language benchmarks don’t measure AI models’ gene... by mikkom on Monday August 24, 2020 @03:00AM (#24258131)

> An interesting example would be if the NN learns something from a book, also fed it with the all the critique on that book ( the environment reaction ).

But isn't that exactly what has been done with GPT-3? AFAIK it has been trained with both facts (wikipedia and such) and web-crawled text content that contains both content and critique of content (book reviews, movie reviews etc.)?

[ Parent | on HN ]

Re:Natural language benchmarks don’t measure AI models’ gene... by ccozan on Monday August 24, 2020 @08:35AM (#24259990)

I know that, but does the network perceive the critique as environmental reaction? It means if it can distinguish between them.

Putting them in the same basket not sure if it helps.

[ Parent | on HN ]