The tokens can be variable length too. If "911" appears often enough in the data used to make the BPE encoding, it may get assigned its own token. If it does, t

The tokens can be variable length too. If "911" appears often enough in the data used to make the BPE encoding, it may get assigned its own token. If it does, the model has no way to "pick it apart" and learn that it actually constructed of a 9, a 1 and a 1. That information has been thrown away, and must be laboriously re-learned from context during training, for those contexts where it's relevant (For instance, if it's part of a math task, or you need a word to rhyme with "been shot down" in a song in the style of Wyclef Jean). It's quite possible that this trade-off is worth it, that making math and poetry harder to learn isn't a big deal because of the savings elsewhere (models can still learn it eventually, as we have seen). But BPE isn't exactly elegant. It's itself a learned encoding, trained on a different objective from the rest of the system. In the rest of the training, we let the GPT model decide how to assemble lower-level concepts into higher-level ones, and which information to ignore. On the lowest level we've hardcoded it for the model, and we know the model sometimes have to work hard to undo our encoding.