Hacker News Reader

If you rewind my brain I expect to give you a similar answer to the one I gave you before. Which isn’t the case for LLM. You can literally replay a positive answer, then get a negative response that isn’t consistent at all with the one it previously generated.

View on HN

I'd argue it is mostly a technical constraint in some LLMs which is due to a few optimization factor (injected temperature, random rounding error caused by parallelism). In practice you could very well create a LLM that always reply the same thing for the same input, but it would take more time to complete (to be sure that the operations are made in the same order). I don't think those ones would differ so much from the "random ones" to call the firsts conscious and the second non-conscious.

That's because you aren't actually rewinding. You're replaying the conversation you just had through the LLM and it's giving you a likely explanation for what it might have said.

It is actually possible to rewind LLMs and get the same response, but it's not typically done both as an optimization and as a defense against distillation.

Reply on news.ycombinator.com