Whats funny about LLMs is they are trained from books, but they are also trained to not output books, so how much of an LLM skews its output because a perfectly normal sentence could be a quote in like 300 different books?
Whats funny about LLMs is they are trained from books, but they are also trained to not output books, so how much of an LLM skews its output because a perfectly normal sentence could be a quote in like 300 different books?