Troed Sångberg on Nostr: Shannon's theorem is enough to show that LLMs can not be explained from them storing ...Shannon's theorem is enough to show that LLMs can not be explained from them storing the training data.
