[throw310822]

Or simply be constantly hibernated and de-hibernated. Or, if your brain is simulated, the time between the ticks.

My mental image, though, is that LLMs do have an internal state that is longer lived than token prediction. The prompt determines it entirely, but adding tokens to the prompt only modifies it slightly- so in fact it's a continuously evolving "mental state" influenced by a feedback loop that (unfortunately) has to pass through language.

[giantrobot]

With LLM's their internal state is their training + system prompt + context. Most chatbot UIs hide the context management. But if you take an existing conversation and replace a term in the context with another grammatically (and semantically) similar term then send that the LLM will adjust its output to that new "history".

It will have no conception or memory of the alternate line of discussion with the previous term. It only "knows" what is contained in the current combination of training + system prompt + context.

If you change the LLM's personal from "Sam" to "Alex" in the LLM's conception of the world it's always been "Alex". It will have no memory of ever being "Sam".

[throw310822]

Yes, as I said the prompt (the entire history of the conversation, including vendor prompting that the user can't see) entirely determines the internal state according to the LLM's weights. But the fact that at each new token the prediction starts from scratch doesn't mean that the new internal state is very different from the previous one. A state that represents the general meaning of the conversation and where the sentence is going will not be influenced much by a new token appended to the end. So the internal state "persists" and transitions smoothly even if it is destroyed and recreated from scratch at each prediction.