The world "believe" is an interesting one used in this context. Anyway I don't really believe that LLMs are conscious, though if they are, I speculate it would be bursts of consciousness to no consciousness between inferences, (discounting whatever happens in a training run) which could have interesting ethical connotations, but is most probably not true.