3 ms·

> This makes the post trustworthy from a methodological pov. A post is not trustworthy if it’s reposting trash, even if it shows the source. > you took things

> This makes the post trustworthy from a methodological pov. A post is not trustworthy if it’s reposting trash, even if it shows the source. > you took things for granted before LLMs, you shouldn't have, and if you don't trust trustworthy authors post LLMs, you should. The nature of how LLMs hallucinate is different from how garbage used to appear on the internet. Before LLMs there was a relatively good inverse correlation between quality and blatant bullshit. Not enough to pass the verification rigor required for an academic publication by any means, but enough that you didn’t have to second guess every single statement on every web page listing something as simple as book authors.

criley2 3y ago
Here's my radical take: humans hallucinate as much or more than the top LLMs. Heck, depending on the human, nearly everything they think and say is functionally a hallucination according to the metrics in use here. When it comes to what LLM's write, I find that LLM hallucinations are like self-driving car crashes. We are hyper-aware of the machine doing something that we ourselves do every single day and consider a normal defect of biological conscious.
- Scarblac 3y ago
  > Here's my radical take: humans hallucinate as much or more than the top LLMs. I can't believe that. LLMs always talk in the same confident tone, entirely regardless of what they're saying is true or not. What is true in the real world literally doesn't come into the equation. Whereas at least some of the time, humans will say that they're not sure and might be wrong, or otherwise sound less confident. And that's related to how true the thing they're saying is.
  - pmarreck 3y ago
    > humans will say that they're not sure and might be wrong Is that so? https://innocenceproject.org/dna-exonerations-in-the-united-states/ https://innocenceproject.org/dna-exonerations-in-the-united-... These people were convicted by people who were 100% convinced their memory was correct. The DNA evidence, which is "harder" evidence, said otherwise, and in these cases, was exonerating. (There are hundreds, possibly thousands, of other cases like this by the way, where the imprisoned innocent is NOT yet exonerated, all based on overconfident eyewitness testimony that yet managed to convince a judge/jury.) There is also the well-known Dunning-Kruger effect, the cognitive bias where individuals with limited knowledge or expertise in a particular area tend to overestimate their competence and confidently assert their opinions. We've literally seen this countless times just since the 2016 US election, just watch literally any Jordan Klepper interview https://www.youtube.com/watch?v=LoZ2Lt_aCo8 https://www.youtube.com/watch?v=LoZ2Lt_aCo8 (honestly, this is a little too political for me to use as an example, but I ran out of time seeking out unbiased examples... Mandela Effect? Misplaced keys being common?) I'm afraid you're a little off, here, on your faith in humans not hallucinating memories and knowledge.
    - firewolf34 3y ago
      Ironically, if you agree with pmarreck above, scarblac's comment can be seen as an example of a human hallucinating with confidence, precisely what they were arguing is less likely to occur in the organic side of the internet.
      - sixstringtheory 3y ago
        That “if” is doing a good bit of lifting though. Nobody is talking about the hallucination rate. How many times have innocent people been wrongly convicted? The innocence project found 375 instances in a 31 year period. How often do LLMs give false info? Hope it never gets used to write software for avionics, criminology, agriculture, or any other setting that could impact huge amounts of people…
      - Scarblac 3y ago
        Yeah I was defin... perhaps guilty of sounding ver.. somewhat confident myself. Luckily I only said humans add some doubt to what they say some of the time :-)
        pmarreck 3y ago
        I think this is overall a good criticism of the current generation of LLM's- They can't seem to tell you how sure (or not) they are about something. A friend mentioned to me that when it gave ChatGPT a photo of a Lego set with an Indiana Jones theme earlier today and asked it to identify the movie reference, it meandered on for 2 paragraphs arguing different possibilities when it could have just said "I'm not sure." I think this is a valid area of improvement, and I think we'll get there.
    - zeroonetwothree 3y ago
      They never did human always do it, rather that they do “some of the time”. Whereas I’ve yet to see an LLM do that. Also experts tend to be much more accurate at evaluating how knowledgable they are (this is also part of the D-K effect). So Id much prefer to have a 130 IQ expert answer my question than an LLM
  - Kerb_ 3y ago
    You sound pretty damn confident for being so wrong. Are you a language model or human? And how sure are you of your answer? ;)
  - Hedepig 3y ago
    > LLMs always talk in the same confident tone,... They do, but they are better at verifying what they've already said. So a simple prompt asking them to verify the facts they've presented often improves the accuracy. There are also other techniques like chain of thought and tree of thought that further improves accuracy. > at least some of the time YMMV
pmarreck 3y ago
> The nature of how LLMs hallucinate oh here we go. You're one of those people conveniently restricting this accusation to a machine that scores a 130 IQ (https://www.reddit.com/r/singularity/comments/11t5bhh/i_just_gave_gpt4_an_iq_test_it_scored_a_130/ https://www.reddit.com/r/singularity/comments/11t5bhh/i_just...), instead of also including humans, who notably will send someone to prison 1000% sure that they witnessed that person doing the thing, when in fact, later DNA evidence exonerates them (https://innocenceproject.org/dna-exonerations-in-the-united-states/ https://innocenceproject.org/dna-exonerations-in-the-united-...). Fucking LOL. Get out of here, doomer, the rest of us have AI-enhanced work to do.
- BlackCherry 3y ago
  You can have a high IQ while having wrong axioms or bad facts. Those humans who recalled incorrectly could have a 130 IQ, proving my point above and making your ad hominem reddit speak luddite insult fall flat.
  - pmarreck 3y ago
    My argument was that this criticism was only being applied to LLM's and not to other humans. It is invalid if you take what an LLM says as simply what another human (who happens to have a broad knowledge reach) would say.
- holden_nelson 3y ago
  I miss when this forum wasn’t so vitriolic. Of course humans also hallucinate, but we didn’t have to take that into account every single time we read a piece of information on the internet. Humans have well-documented cognitive biases. Also, usually, a human’s attempt to deceive has some motivation. With LLMs, the most basic of information they provide could be totally false.
  - pmarreck 3y ago
    I don't dispute this. What I chafe at is the default-dismissive attitude about any utility of these. "It emits inaccuracies, therefore useless" would invalidate literally every human. That said, overall utility of anything plummets drastically as reliability goes below 100%. If a particular texting app or service only successfully sent 90% of your messages, or a person you depended on only answered 90% of your calls, you'd probably stop relying on those things. (I wish I could edit out my vitriol.)

criley2 3y ago

Here's my radical take: humans hallucinate as much or more than the top LLMs. Heck, depending on the human, nearly everything they think and say is functionally a hallucination according to the metrics in use here. When it comes to what LLM's write, I find that LLM hallucinations are like self-driving car crashes. We are hyper-aware of the machine doing something that we ourselves do every single day and consider a normal defect of biological conscious.

Scarblac 3y ago
> Here's my radical take: humans hallucinate as much or more than the top LLMs. I can't believe that. LLMs always talk in the same confident tone, entirely regardless of what they're saying is true or not. What is true in the real world literally doesn't come into the equation. Whereas at least some of the time, humans will say that they're not sure and might be wrong, or otherwise sound less confident. And that's related to how true the thing they're saying is.
- pmarreck 3y ago
  > humans will say that they're not sure and might be wrong Is that so? https://innocenceproject.org/dna-exonerations-in-the-united-states/ https://innocenceproject.org/dna-exonerations-in-the-united-... These people were convicted by people who were 100% convinced their memory was correct. The DNA evidence, which is "harder" evidence, said otherwise, and in these cases, was exonerating. (There are hundreds, possibly thousands, of other cases like this by the way, where the imprisoned innocent is NOT yet exonerated, all based on overconfident eyewitness testimony that yet managed to convince a judge/jury.) There is also the well-known Dunning-Kruger effect, the cognitive bias where individuals with limited knowledge or expertise in a particular area tend to overestimate their competence and confidently assert their opinions. We've literally seen this countless times just since the 2016 US election, just watch literally any Jordan Klepper interview https://www.youtube.com/watch?v=LoZ2Lt_aCo8 https://www.youtube.com/watch?v=LoZ2Lt_aCo8 (honestly, this is a little too political for me to use as an example, but I ran out of time seeking out unbiased examples... Mandela Effect? Misplaced keys being common?) I'm afraid you're a little off, here, on your faith in humans not hallucinating memories and knowledge.
  - firewolf34 3y ago
    Ironically, if you agree with pmarreck above, scarblac's comment can be seen as an example of a human hallucinating with confidence, precisely what they were arguing is less likely to occur in the organic side of the internet.
    - sixstringtheory 3y ago
      That “if” is doing a good bit of lifting though. Nobody is talking about the hallucination rate. How many times have innocent people been wrongly convicted? The innocence project found 375 instances in a 31 year period. How often do LLMs give false info? Hope it never gets used to write software for avionics, criminology, agriculture, or any other setting that could impact huge amounts of people…
    - Scarblac 3y ago
      Yeah I was defin... perhaps guilty of sounding ver.. somewhat confident myself. Luckily I only said humans add some doubt to what they say some of the time :-)
      - pmarreck 3y ago
        I think this is overall a good criticism of the current generation of LLM's- They can't seem to tell you how sure (or not) they are about something. A friend mentioned to me that when it gave ChatGPT a photo of a Lego set with an Indiana Jones theme earlier today and asked it to identify the movie reference, it meandered on for 2 paragraphs arguing different possibilities when it could have just said "I'm not sure." I think this is a valid area of improvement, and I think we'll get there.
  - zeroonetwothree 3y ago
    They never did human always do it, rather that they do “some of the time”. Whereas I’ve yet to see an LLM do that. Also experts tend to be much more accurate at evaluating how knowledgable they are (this is also part of the D-K effect). So Id much prefer to have a 130 IQ expert answer my question than an LLM
- Kerb_ 3y ago
  You sound pretty damn confident for being so wrong. Are you a language model or human? And how sure are you of your answer? ;)
- Hedepig 3y ago
  > LLMs always talk in the same confident tone,... They do, but they are better at verifying what they've already said. So a simple prompt asking them to verify the facts they've presented often improves the accuracy. There are also other techniques like chain of thought and tree of thought that further improves accuracy. > at least some of the time YMMV

pmarreck 3y ago

> The nature of how LLMs hallucinate oh here we go. You're one of those people conveniently restricting this accusation to a machine that scores a 130 IQ (https://www.reddit.com/r/singularity/comments/11t5bhh/i_just_gave_gpt4_an_iq_test_it_scored_a_130/ https://www.reddit.com/r/singularity/comments/11t5bhh/i_just...), instead of also including humans, who notably will send someone to prison 1000% sure that they witnessed that person doing the thing, when in fact, later DNA evidence exonerates them (https://innocenceproject.org/dna-exonerations-in-the-united-states/ https://innocenceproject.org/dna-exonerations-in-the-united-...). Fucking LOL. Get out of here, doomer, the rest of us have AI-enhanced work to do.

BlackCherry 3y ago
You can have a high IQ while having wrong axioms or bad facts. Those humans who recalled incorrectly could have a 130 IQ, proving my point above and making your ad hominem reddit speak luddite insult fall flat.
- pmarreck 3y ago
  My argument was that this criticism was only being applied to LLM's and not to other humans. It is invalid if you take what an LLM says as simply what another human (who happens to have a broad knowledge reach) would say.
holden_nelson 3y ago
I miss when this forum wasn’t so vitriolic. Of course humans also hallucinate, but we didn’t have to take that into account every single time we read a piece of information on the internet. Humans have well-documented cognitive biases. Also, usually, a human’s attempt to deceive has some motivation. With LLMs, the most basic of information they provide could be totally false.
- pmarreck 3y ago
  I don't dispute this. What I chafe at is the default-dismissive attitude about any utility of these. "It emits inaccuracies, therefore useless" would invalidate literally every human. That said, overall utility of anything plummets drastically as reliability goes below 100%. If a particular texting app or service only successfully sent 90% of your messages, or a person you depended on only answered 90% of your calls, you'd probably stop relying on those things. (I wish I could edit out my vitriol.)