It would be like logging a bunch of random noise from anyone's perspective except the LLM's.
I guess I'm also just confused. I get that this is _difficult_ to do, but I would think that computer scientists would be utterly dissatisfied that AI was "non-deterministic" and would poke at the problem until it could be understood.
I guess I'm also just confused. I get that this is _difficult_ to do, but I would think that computer scientists would be utterly dissatisfied that AI was "non-deterministic" and would poke at the problem until it could be understood.