3 ms·
>Perhaps it is able to tell, if you ask it? Ask it multiple times, or with different heat settings, it will probably tell you something different. Tell it you
by krapp 3y ago
>Perhaps it is able to tell, if you ask it?
Ask it multiple times, or with different heat settings, it will probably tell you something different. Tell it you own Star Wars and it will respond in kind. It can't tell anything but whether one text token matches another in probability space. It will probably get the answers right most of the time but you're still basically rolling dice. Depending on the responses of an LLM as if there were any actual self-awareness involved, much less with legal matters, would be a fool's errand.
- danielbln 3y agoThis argument only works if you assume all output of an LLM comes merely from its training data, and that it receives no alignment via RFHL, no outside data ground truth via RAG and so on. The engine might be a probabilistic token predictor, but the car is the sum of its part and those parts are not just the engine.
- krapp 3y agoNo it still works, because in that case all the LLM is doing is making an API call based on the same token prediction, and passing on the result. It still doesn't "know" anything about anything.