Jayhem

This is the most compelling articulation of what is going on in autoregressive models [paraphrased, by @ilyasut ] Internet texts contain a projection of the world. Learning the underlying dynamics that created this text is the most efficient way of doing next token prediction - therefore, pre-training constructs a world model RLHF et al then imbue the model with intentions, or the knowledge of how to navigate this world model to arrive at a desired outcome