Comment by BUFU
Would it be possible that other people posted content of Harry Potter book online and the model developer scrape that information? Would the model developer be at fault in this scenario?
Would it be possible that other people posted content of Harry Potter book online and the model developer scrape that information? Would the model developer be at fault in this scenario?
I think this is good question. At least for LLMs in general. However we know that Meta used pirated torrents.