A classically pretrained LLM does not have a concept of having determined its previous tokens — it has only ever observed inputs that it had no causal influence

A classically pretrained LLM does not have a concept of having determined its previous tokens — it has only ever observed inputs that it had no causal influence over. This is why it's valid to say its actions are predictive and not determinative.