The model isn’t sampling from training data during inference. It’s sampling from a latent space created during training, which used all of the training data to

The model isn’t sampling from training data during inference. It’s sampling from a latent space created during training, which used all of the training data to create it. Fundamentally, the presence of a next token in some of the samples in the corpus is just as informative as the absence of that token from other samples. You can’t cite one without the other. If you wanted to give attribution you’d need to list the entire training corpus.