Many companies stopped publishing their data sets after people published evidence they were mass, copyright infringement. They dropped the specifics of pretrain

Many companies stopped publishing their data sets after people published evidence they were mass, copyright infringement. They dropped the specifics of pretraining data from the model cards. Aside from licensing content, that content creators don’t like redistribution means a lawful model would probably only use Gutenberg’s collection and permissive code. Anything else, including Wikipedia, usually has licensing requirements they might violate.