I suspect the challenge is avoiding cases where answers are overfit. Can you get it to give you the full text of a chapter in a book that it has read, for examp

I suspect the challenge is avoiding cases where answers are overfit. Can you get it to give you the full text of a chapter in a book that it has read, for example? It'd be easy to imagine models that accidentally made this possible.