Gemini, gpt and fable are actually very good compressions of internet content. But is lossless compression as in they kept the most important part (for them to fulfill the next token task) and found a way to mimic the rest.
1 comments
I think you meant lossy compression and not lossless. I'm not suggesting this as a method to extract those books from the models, which by their nature are not databases. Just commenting on the somewhat surprising fact that the bigger the model the more likely it is to produce some (short) excerpts of the original training material