Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI hallucinations: Why LLMs make things up (and how to fix it)

kapa.ai

Post reply on HN
Print view View original

251–257 of 257 posts

Re: AI hallucinations: Why LLMs make things up (and how to fix it)

#251

Post by TZubiri »

Gormo wrote: post #158
TZubiri wrote: post #94

Earlier quoted context omitted.

"I don't think that definition works: it's attempting to categorize statements according to criteria completely external to them rather than according to any inherent property of the statement." Correct. The basic concept of truth in logic relies on an objective reality, an expression a priori holds truth even in the absence or indistinct of such a reality. But the truthfulness or correctness of a posteriori statemen…

> Correct. The basic concept of truth in logic relies on an objective reality, an expression a priori holds truth even in the absence or indistinct of such a reality. But the truthfulness or correctness of a posteriori statements can depend on the reality. Examples of the former would be "If A is B, then B is C. A is B, then B is C" Example of the latter would be "It is raining outside." What you're describing is the…

"We're talking about specific outputs generated by the LLM, not the LLM itself. The training data consists of prior expressions of language which in turn may be influenced by human observations of reality, but the LLM is only ever making probabilistic inferences based on that second-order data"

You recognize that training data are influenced by human observations. And that LLM outputs are influenced by training data (and fine tuning). So it follows that LLM outputs are influenced by observations of the world. Why would the causality chain stop after 2 links?

https://chatgpt.com/share/67534483-8e6c-800f-9534-d764a90981...

You may call this a hallucination, but it is for sure based on observation. Otherwise the LLM wouldn't know the answer. It is undeniable that LLMs have empirical knowledge of the world through embedded human observation.

Re: AI hallucinations: Why LLMs make things up (and how to fix it)

#252

Post by __MatrixMan__ »

namaria wrote: post #131
Timwi wrote: post #129

Earlier quoted context omitted.

Of course they can. Carbon dioxide consists of quarks and electrons. I can divide it into units smaller than atoms and it's still quarks and electrons. All you did was a word trick by assuming a specific meaning of “substance”.

No word trick. Just pointing out that there's some nuance to it. > Carbon dioxide consists of quarks and electrons. But this is just plain wrong. Carbon dioxide consists of carbon dioxide molecules. There is no "carbon-dioxidity" to the quarks and electrons (which are also made of quarks) that the atoms that make the molecules can be broken down into.

That's a good point, I hadn't thought about it that way. Maybe it's a little less of an arbitrary choice than I was giving it credit for.

It is not, however, Democratus's point. So I think the "words change, deal with it" argument still stands.

Re: AI hallucinations: Why LLMs make things up (and how to fix it)

#253

Post by fsckboy »

mdp2021 wrote: post #124
fsckboy wrote: post #92

it's superficially counterintuitive to people that an AI that will sometimes spit out verbatim copies of written texts, also will just make other things up. It's like "choose one, please". MetaAI makes up stuff reliably. You'd think it would be an ace at baseball stats for example, but "what teams did so-and-so play for", you absolutely must check the results yourself.

> "counterintuitive" It is consistent with the topic that the reply would be "Tell them that sequences of words that were verbatim in a past input have high probability, and gaps in sequences compete in probability". Which fixes intuition, as duly. In fact, things are not supposed to reply through intuition, but through vetted intuition (and "vetted mature intuition", in a loop). > you absolutely must check the resul…

i said "superficially counterintuitive", you misquoted me and proceeded with a non superficial comment.

Re: AI hallucinations: Why LLMs make things up (and how to fix it)

#254

Post by mdp2021 »

fsckboy wrote: post #253
mdp2021 wrote: post #124

Earlier quoted context omitted.

> "counterintuitive" It is consistent with the topic that the reply would be "Tell them that sequences of words that were verbatim in a past input have high probability, and gaps in sequences compete in probability". Which fixes intuition, as duly. In fact, things are not supposed to reply through intuition, but through vetted intuition (and "vetted mature intuition", in a loop). > you absolutely must check the resul…

i said "superficially counterintuitive", you misquoted me and proceeded with a non superficial comment.

> misquoted me

Why? The reply would have been the same if I quoted the whole «superficially counterintuitive to people that [...]» (instead of just pointing to the original).

> proceeded with a non superficial comment

Well, hopefully ;)

Your post went into the right direction of leading towards the idea that "there is intuition, and there is mature thought further from that: and processors must not stop at intuition, immature thought".

(Stochastic output falls in said category of "intuition"... As "bad intuition", since it goes in the wrong direction in the vector "naive to sophisticated".)

Re: AI hallucinations: Why LLMs make things up (and how to fix it)

#255

Post by watwut »

PittleyDunkin wrote: post #205
watwut wrote: post #201

Earlier quoted context omitted.

Expected defects are bugs too. I totally expect half the problems in the software my company is developing. They are still bugs.

What is the utility of this sense of "bug"? If not all bugs can be fixed it seems better to toss the entire concept of a "bug" out the window as no longer useful for describing the behavior of software.

What is utility of any other sense? I expect null pointer to happen. It is still a bug. Even if it is in some kind of special situation we dont have time to fix.

> If not all bugs can be fixed it seems better to toss the entire concept of a "bug" out the window as no longer useful for describing the behavior of software.

Then those are bugs you cant fix. It is just lying to yourself to call them not a bug ... if they are bugs.

Re: AI hallucinations: Why LLMs make things up (and how to fix it)

#256

Post by sam_1m »

tshadley wrote: post #174
sfink wrote: post #37

Earlier quoted context omitted.

> proving that hallucination is conceptually a simple problem. ...proving that this one particular piece of the hallucination problem may be conceptually simple. FTFY

> ...proving that this one particular piece of the hallucination problem may be conceptually simple. Everything mentioned in the article boils down to that one particular piece-- non-detected uncertainty. The architecture constraints referenced are all situations that cause uncertainty. Training data gaps of course increase uncertainty. Their solutions are a shotgun blast of heuristics that all focus on reducing unce…

[dead]

Re: AI hallucinations: Why LLMs make things up (and how to fix it)

#257

Post by stonemetal12 »

SirMaster wrote: post #223
stonemetal12 wrote: post #207

Earlier quoted context omitted.

> LLMs naturally hallucinate, but it is not what we want, so it is a bug. I rolled a one in D&D, it is not what I wanted, so it is a bug. Remove it from all my dice.

What? You are telling me that when you roll a 6 sided dice you are not expecting any of the 1-6 as a result? If a 6-sided dice produced a 7 that would be a bug. When you rolled a dice, I would argue that you knew you wanted a random number from 1-6, not that you wanted a specific number or not a specific number. If you wanted that you wouldn't have used a dice. When I ask an LLM to write code for me and it references…

>You are telling me that when you roll a 6 sided dice you are not expecting any of the 1-6 as a result?

The statement I replied to wasn't any non-expected result is a bug, it was non-desired output is a bug (hence the joke about not desiring an expected output). LLMs producing "funny" (hallucination) outputs are expected but only sometimes not desired, therefore not a bug in my opinion.

How do you use an LLM in story telling if it isn't allowed to produce fictious outputs?

Post reply on HN