6 ms·
LLMs _only_ produce hallucinations. Some of them align with the someone's version of the truth (assuming that is what the user wanted)
by 4ndrewl 3y ago
LLMs _only_ produce hallucinations. Some of them align with the someone's version of the truth (assuming that is what the user wanted)
- consumer451 3y agoI look forward to the inevitable work of science fiction titled Latent Space.
- optimalsolver 3y ago"Engines won't respond, commander. We're stuck in a latent space polarity vortex."
- moffkalast 3y ago"Auxilary power to the deflector, a modulated antiproton pulse might set us free."
- rexf 3y ago"Hallucinations" is too charitable of a term for making stuff up. It personifies "AI" which doesn't need help selling its "intelligence".
- _nalply 3y agoYou could use "confabulate" instead, could you? "Make stuff up" is very human and would personify "AI" even more.
- dragonwriter 3y agoConfabulations are also very human, but a much better metaphor than hallucination for the actual errors LLMs make.
- dragonwriter 3y agoHallucinations are false current sense perceptions. LLMs don’t have senses and don’t hallucinate at all; the LLM errors described as “hallucinations” are closer, if one needs an anthropomorphizing metaphor, to confabulations. Which kind of make sense; as LLMs have almost no memory, just an instinct to respond and some instinctual responses (the result of “training”, which is also a bad metaphor; only “in-context learning” is analogous to training/learning for humans, what is called “training” is guided evolution of frozen instincts) and whatever is in their context window. And lack of memory plus a prompt to respond is a major context where confabulations happen with humans (these are specifically called “provoked confabulations.”)
- hallqv 3y agoYou must have a very peculiar definition of a hallucination. Why would an LLM reciting a fact correctly be hallucinating?
- smeagull 3y agoSuppose I make a massive book of predictions. Some of which turn out to be correct. Am I now capable of predicting the future? Suppose I wrote the book to be as banal (i.e. highly probable) as possible. Am I predicting the future now? And, how impressive is it?
- M4v3R 3y ago> Suppose I make a massive book of predictions. Some of which turn out to be correct. Am I now capable of predicting the future? If you write a book of random predictions without any insight the vast majority of them will be false, so even if few of them are right it is not impressive nor anyone would say you're capable of predicting the future. In comparison, the OP states that GPT-4 predictions are 97% correct. And yes, I would say that is pretty impressive. If 97% of anything I say about the future was correct I would be considered a wizard and probably be a billionaire.
- vsnf 3y ago>If you write a book of random predictions without any insight the vast majority of them will be false, so even if few of them are right it is not impressive nor anyone would say you're capable of predicting the future. Isn't this step for step exactly what Nostradamus did?
- M4v3R 3y agoBasically yes, plus he was as vague as he could be to cover many possible outcomes. That’s why no educated person takes his „prophecies” seriously.
- numpad0 3y ago> If you write a book of random predictions without any insight the vast majority of them will be false And you just hallucinated that! It's just, if there was a system that only talks in an oddly definitive hypotheticals, they can be correct about a lot, and GPT-* are exactly that.
- YetAnotherNick 3y agoThis has been 3rd or 4th time in HN I have commented on top voted posts making very general statement about LLM without defining the terms. And I have never gotten any reply. LLM discussions in HN has become so shallow like there are two sides. Please define what you mean by hallucination.
- tveita 3y agoA different way to state the claim may be "whatever LLMs do, they do it all the time" There are no separate modes of operation when GPT-4 is telling me about Javascript's "console.log" method, and when it's telling me about Rackets non-existing "sum" function. The network has no way of telling that the strands of probability are wearing thin under it's feet. The engine spins and churns out another token and we deem it "correct" or "incorrect".
- TerrifiedMouse 3y agoHallucination in this context is when LLMs make up stuff. The OP's assertion is "make up stuff" is all LLMs really do and sometimes the stuff they made up matches reality - and sometimes it doesn't.
- intended 3y ago“the generated content that is nonsensical or unfaithful to the provided source content” Ji et al 2022
- YetAnotherNick 3y agoIf I go with this definition, and ask it say what is 2*2, it is not nonsensical or unfaithful. So there is one example where it is not hallucinating?
- 4ndrewl 3y agoOr where hallucinating truth co-indices with our understanding of truth?
- 3y ago
- smeagull 3y agoTruth is an entirely orthogonal concept to the data that they have been trained on. Sometimes LLMs might approach coherency.
- _nalply 3y agoI agree with this comment. In fact I would even stop talking about the profound concept of truth. Long time ago I read this wisdom: «God's omnipotence raises the intriguing possibility that two contradictory religions could both be true.» Buddha says anyway, that everything is empty. Zen Buddhism employs koans to dismiss truth. LLMs sometimes hit the mark and this makes many people ask what is truth and what are humans?
- dudeinjapan 3y agoMost religions started as hallucinations.
- RandomLensman 3y agoThere are facts (aka truth), for example, as to what sentences or words are actually in a document and which ones are not. Having LLMs evolve us into complete incompetence as to basic facts is not the way forward.
- hospitalJail 3y ago>this makes many people ask what is truth and what are humans? After enough philosophy, this is close to a solved problem. There are so many breakdowns of what is truth, including adding words to describe different truths. If you are hung up on a single universal definition, you are most likely limited by language, a human construct rather than the understanding of Truth. We all know what Truth is, we know that 2+2=4. As Laozi would say 'If you try to grasp it, it will slip". There is a time and a place for analysis, but there is also losing the forest for the trees.
- column 3y agoPardon my French but this is the exact same of bullshit pseudo-philosophical semantic non-arguments we hear from jordan peterson (please watch the Some More News video about him). "What even are words?"
- 4ndrewl 3y agoThanks. I'm more of a reader than a video viewer though. It's really around the epistemological limits of LLMs, and truth. Probably more Kant than Peterson?
- amelius 3y ago"Some" is increasing with every new generation of LLMs, though. And there is no reason to believe it won't converge to 100%.
- weinzierl 3y agoThis matches my view, but it begs the question what exactly is measured by "Halucination Rate"?
- Alifatisk 3y ago> LLMs _only_ produce hallucinations. Some of them align with the someone's version of the truth I partly disagree because not all truth is subjective, but rather, objectively true.
- intended 3y agoHeck yes. We are so used to assuming that good text means good thought. "emergent" behavior is assumed to be full on thinking and learning systems. Calling it Hallucination is absurd. Humans hallucinate. Humans can be deceived or deceive because there is an active effort to hide something. An LLM is producing the next best token. Its either always hallucinating or never hallucinating.