Skip to content

Comment on Ask HN: Anyone else disillusioned with "AI experts" in their team?parent

Comments

Okay yes, but would you really say that the main part of non-determinism in LLM-usage stems from this ? No its obviously the topk sampling.

I don't think my tech-lead was trying to suggest the floating-point error/non-associativity was the real source.

Would you really say that the main part of non-determinism in LLM-usage stems from this

Yes I would because it causes exponential divergence (P(correct) = (1-e)^n) and doesn't have a widely adopted solution. The major labs have very expensive researchers focused on this specific problem.

There is a paper from Thinking Machines from September around Batch Invariant kernels you should read, it's a good primer on this issue of non-determinism in LLM's, you might learn something from it!

Unfortunately the method has quite a lot of overhead, but promising research all the same.

Alright fair enough.

I dont think this is relevant to the main-point, but it's definitely something I wasn't aware of. I would've thought it might have an impact on like O(100)th token in some negligible way, but glad to learn.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.