The combination of the most probable tokens doesn't necessarily have to be in th...

e12e · 2026-07-26T16:33:13

If you shuffle a deck of cards, you likely have a novel sequence of cards. I'm not sure how that is relevant?

josu · 2026-07-26T21:38:13

I was responding to this "merely produces syntactically likely combinations of the training data".

DiabloD3 · 2026-07-26T09:01:27

That is called a hallucination.

josu · 2026-07-26T09:46:11

Yes, like the Jacobian counterexample.

inigyou · 2026-07-26T11:29:33

That LLM was attached to a Python and MATLAB evaluator. It generated plausible things to feed into a real computer.

pixl97 · 2026-07-26T17:10:52

Weird take as a lot of things humans invent and test are done in similar ways.

Gander5739 · 2026-07-26T09:12:48

Only if it's incorrect.