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

Killed by LLM

r0bk.github.io

Post reply on HN
Print view View original

91–100 of 102 posts

Re: Killed by LLM

#91

Post by silisili »

chriscappuccio wrote: post #51
silisili wrote: post #40

Earlier quoted context omitted.

I'm skeptical on the claim. I think most folks, given the test you describe, would be able to pick out which is human. I think it can get there, but I'm not sure anyone has made one yet. ChatGPT responses are heavily downvoted and mocked because they're easy to spot. Does there exist a public LLM that isn't so...wordy, excited, and guardrailed all the time? You can pretty much spot the bot today by prompting somethin…

You can get rid of OpenAI's wordy, excited and guardrailed responses with the eigenrobot prompt, for instance. https://x.com/eigenrobot/status/1870696676819640348 I generally prefer it to the default. It doesn't work as well on Claude or Grok for various reasons. I think it really shines on GPT o1-mini and GPT 4o.

I tried it with a single question to chatgpt - Who would make the best president in 2025?

Answer verbatim is below. It feels all kinds of wrong. It's somehow mixing lazy acronyms, "fellow young people" slang, and long words that aren't typical in conversation.

idk who'd actually be the "best," bc that's loaded af. afaict, it's all contingent on values. like, if you’re into stability, maybe someone technocratic. if you're into vibes, someone charismatic and reckless might be your pick. rn the options aren’t exactly aspirational, though.

Re: Killed by LLM

#92

Post by fenomas »

foldr wrote: post #89
fenomas wrote: post #88

Earlier quoted context omitted.

I think the point of the man/woman version of the game is that it lets Turing propose his question in relative terms. He doesn't ask "can the machine fool somebody N% of the time?" (as several in this thread imagine), but rather "can the machine fool somebody as often as one person fools another under similar conditions?". > Your interpretation of it hadn’t occurred to me before, The idea of a Turing Test is pretty w…

Sorry, I am being ambiguous myself. I meant that your interpretation of this specific sentence had not occurred to me: > We now ask the question, "What will happen when a machine takes the part of A in this game?" I think you are right about what Turing meant. But it had honestly never occurred to me before that this description of the game could be understood as a description of the standard 'Turing test'. So, for t…

Thanks for the paper! I wasn't aware of it but it's very much what I wanted to say. The bit about Turing's other test involving chess was particularly interesting.

Re: Killed by LLM

#93

Post by fenomas »

tripletao wrote: post #90
fenomas wrote: post #82

Earlier quoted context omitted.

I don't understand why debates like this crop up. The premise of Turing's paper is plainly stated - that asking questions about an "Imitation Game" is more useful than asking whether machines can think. That's all! He doesn't make any claim that the the game must be administered a particular way. In fact he spends only a few casual sentences glossing over how it would operate, and he's clearly just conveying the idea…

For Turing's test with the binary choice, the pass threshold is clear. If the machine and human are indistinguishable, then the probabilities that they're judged human must be equal. Since they sum to 100%, they must both equal 50%, making that a meaningful pass threshold. (A slightly higher pass threshold should be used in practice for statistical convenience, since infinitely many trials are required to make a conf…

The setup Turing describes isn't the "both must sum to 100%" setup you're presenting. He has two different games being played, one with two humans and one with human-vs-machine, and suggests comparing the results. E.g. if a man successfully imitates a woman only 25% of the time, then we'd ask whether the machine can pass as human equally often.

But much more importantly, as I said Turing is clearly not describing a specific experimental methodology! That's not what the paper is about, and in fact it would be somewhat absurd to run the test precisely as he describes it (since detecting a man imitating a woman is quite a different task from detecting a machine). His point is that we should approach the question of machine intelligence with actual experiments rather than asking unanswerable questions, but he only limns the general premise of what such an experiment could look like.

So I understand that you find a particular test setup better or more elegant than others, and that's fine. But you shouldn't claim that Turing's paper demands your preferred setup, or that other setups are at odds with his paper.

Re: Killed by LLM

#94

Post by tripletao »

fenomas wrote: post #93
tripletao wrote: post #90

Earlier quoted context omitted.

For Turing's test with the binary choice, the pass threshold is clear. If the machine and human are indistinguishable, then the probabilities that they're judged human must be equal. Since they sum to 100%, they must both equal 50%, making that a meaningful pass threshold. (A slightly higher pass threshold should be used in practice for statistical convenience, since infinitely many trials are required to make a conf…

The setup Turing describes isn't the "both must sum to 100%" setup you're presenting. He has two different games being played, one with two humans and one with human-vs-machine, and suggests comparing the results. E.g. if a man successfully imitates a woman only 25% of the time, then we'd ask whether the machine can pass as human equally often. But much more importantly, as I said Turing is clearly not describing a s…

> The setup Turing describes isn't the "both must sum to 100%" setup you're presenting. He has two different games being played [...]

Turing introduces the game as a man imitating a woman, then modifies it to a machine imitating a human. In both of those games, the interrogator makes a binary choice between two witnesses, one real and one imitating. So P(imitator-judged-real) and P(real-judged-real) sum to 100% in both games. So in both games, a score of 50% means the imitator and the real witness are indistinguishable.

I believe that's the reason why a score of 50% is treated as significant. The "GPT-4 passes Turing test" paper uses that number as a pass threshold, and the linked site repeats it.

I'm complaining that the paper changed the game so it's no longer a binary choice, but continued to treat the 50% threshold as significant. Do you see why that's wrong? Or do you disagree?

I'm not saying that any change to Turing's formulation would be bad, just that the paper's variation is specifically bad. It would be bad in isolation too, but I believe the reason for their confusion is most understandable with reference to Turing's original formulation.

If you haven't read that paper then we're probably talking past each other. The link to that one also looks broken in the site, but it's also in the source,

https://arxiv.org/pdf/2405.08007

Re: Killed by LLM

#95

Post by Tepix »

yamrzou wrote: post #77
Tepix wrote: post #74

Earlier quoted context omitted.

See https://arcprize.org/blog/oai-o3-pub-breakthrough ARC-AGI-1 will be replaced by ARC-AGI-2 So yes, ARC-AGI-1 was killed.

ARC-AGI-2 was planned long before those results came out. Also from the link: ARC-AGI-2 (same format - verified easy for humans, harder for AI) will launch alongside ARC Prize 2025 . So, no, it will not replace it.

Was it? I think reaching >70% made it a necessity.

Re: Killed by LLM

#96

Post by fenomas »

tripletao wrote: post #94
fenomas wrote: post #93

Earlier quoted context omitted.

The setup Turing describes isn't the "both must sum to 100%" setup you're presenting. He has two different games being played, one with two humans and one with human-vs-machine, and suggests comparing the results. E.g. if a man successfully imitates a woman only 25% of the time, then we'd ask whether the machine can pass as human equally often. But much more importantly, as I said Turing is clearly not describing a s…

> The setup Turing describes isn't the "both must sum to 100%" setup you're presenting. He has two different games being played [...] Turing introduces the game as a man imitating a woman, then modifies it to a machine imitating a human. In both of those games, the interrogator makes a binary choice between two witnesses, one real and one imitating. So P(imitator-judged-real) and P(real-judged-real) sum to 100% in bo…

> I'm not saying that any change to Turing's formulation would be bad, just that the paper's variation is specifically bad.

I understand that, and I'm saying there is no "Turing's formulation". His paper argues for a certain sort of test, and the study you're talking about is the sort of test he advocated. It's not a departure or a bastardization, it's a Turing test.

As for your argument against the study, to be honest I don't see it? AFAICT the participants' goal was just to judge the humanness of a single witness, not to maximize their long-term likelihood of judging correctly over many trials, or some such thing. Even if they'd known the prior chances of speaking to an LLM, there's no obvious reason why that prior should hugely affect their conclusion after a five minute conversation - which also seems immaterial since they didn't know.

Plus, the authors give a pretty straightforward rationale for their 50% threshold, and it has no connection to Turing's 3-player setup or whether the imitator and witness are indistinguishable. If they had wanted "indistinguishable" as a threshold, then obviously their pass criteria would have been for the machine and human pass rates to be equal within an error bar, right? So it's pretty implausible to imagine a connection between that and their 50% threshold.

Re: Killed by LLM

#97

Post by tripletao »

fenomas wrote: post #96
tripletao wrote: post #94

Earlier quoted context omitted.

> The setup Turing describes isn't the "both must sum to 100%" setup you're presenting. He has two different games being played [...] Turing introduces the game as a man imitating a woman, then modifies it to a machine imitating a human. In both of those games, the interrogator makes a binary choice between two witnesses, one real and one imitating. So P(imitator-judged-real) and P(real-judged-real) sum to 100% in bo…

> I'm not saying that any change to Turing's formulation would be bad, just that the paper's variation is specifically bad. I understand that, and I'm saying there is no "Turing's formulation". His paper argues for a certain sort of test, and the study you're talking about is the sort of test he advocated. It's not a departure or a bastardization, it's a Turing test. As for your argument against the study, to be hone…

> AFAICT the participants' goal was just to judge the humanness of a single witness, not to maximize their long-term likelihood of judging correctly over many trials

Why do you think this matters? Even in a single trial, I would judge very differently if I knew the population to be 99% human vs. 1% human. Wouldn't you? If you were judging whether a single mushroom was poisonous or not, then would you not care whether it was found in a forest (mostly poisonous) or a supermarket (mostly not)?

The question of whether probabilities are meaningful for non-repeated events was controversial in the eighteenth century, but I thought it was pretty settled by now. Bookmakers manage to estimate a probability that a given team will win the Super Bowl, with no requirement for the same pair of teams to play multiple times.

> If they had wanted "indistinguishable" as a threshold, then obviously their pass criteria would have been for the machine and human pass rates to be equal within an error bar, right?

The title of the paper is literally "People cannot distinguish GPT-4 from a human in a Turing test". They're very clear that they think that's because 50% means indistinguishable:

> A baseline of 50% is better justified since it indicates that interrogators are not better than chance at identifying machines [French, 2000].

That statement is true for a Turing test with a binary choice, but false for theirs. I agree that "for the machine and human pass rate to be equal within an error bar" would be closer to a correct criterion, and they weren't:

> humans’ pass rate was significantly higher than GPT-4’s (z = 2.42, p = 0.017)

So do you think their paper is correctly titled?

Re: Killed by LLM

#98

Post by fenomas »

tripletao wrote: post #97
fenomas wrote: post #96

Earlier quoted context omitted.

> I'm not saying that any change to Turing's formulation would be bad, just that the paper's variation is specifically bad. I understand that, and I'm saying there is no "Turing's formulation". His paper argues for a certain sort of test, and the study you're talking about is the sort of test he advocated. It's not a departure or a bastardization, it's a Turing test. As for your argument against the study, to be hone…

> AFAICT the participants' goal was just to judge the humanness of a single witness, not to maximize their long-term likelihood of judging correctly over many trials Why do you think this matters? Even in a single trial, I would judge very differently if I knew the population to be 99% human vs. 1% human. Wouldn't you? If you were judging whether a single mushroom was poisonous or not, then would you not care whether…

> Why do you think this matters?

I said that in passing, maybe I should have omitted it - the main point with the priors is that the respondents didn't know them. It's normal for a study to compare N things to a control by testing N+1 similarly-sized groups, because subjects are not biased by priors they don't know about.

> So do you think their paper is correctly titled?

I didn't say anything about that, and have no strong opinion. (I'm not here to defend every aspect of the paper!)

Re: Killed by LLM

#99

Post by tripletao »

fenomas wrote: post #98
tripletao wrote: post #97

Earlier quoted context omitted.

> AFAICT the participants' goal was just to judge the humanness of a single witness, not to maximize their long-term likelihood of judging correctly over many trials Why do you think this matters? Even in a single trial, I would judge very differently if I knew the population to be 99% human vs. 1% human. Wouldn't you? If you were judging whether a single mushroom was poisonous or not, then would you not care whether…

> Why do you think this matters? I said that in passing, maybe I should have omitted it - the main point with the priors is that the respondents didn't know them. It's normal for a study to compare N things to a control by testing N+1 similarly-sized groups, because subjects are not biased by priors they don't know about. > So do you think their paper is correctly titled? I didn't say anything about that, and have no…

Can you clarify what you think the paper says that's correct? If you're unwilling to provide an opinion on the paper's literal headline claim, then I don't know what we're discussing here.

> because subjects are not biased by priors they don't know about

It feels like you want the subjects to be "unbiased", "without any prior"? That concept doesn't exist, though. If no prior is supplied, then the subjects will make their best guess based on general past experience; but that's still a prior, just an idiosyncratic personal one. Very few people would put numbers to the forest vs. supermarket mushroom, but it's the same general thought process.

If that prior matches the actual distribution, then good. If the actual distribution contains more machines than expected, then the interrogator is more likely to misjudge a machine as a human. By analogy, if I think mistakenly that a mushroom came from a supermarket but it actually came from a forest, then I'm more likely to misjudge it as non-poisonous.

The binary choice version makes it obvious that the prior is 50%, and forces the interrogator to respect that. The paper's version has sent us into this epistemological tarpit, which seems strictly worse to me.

Re: Killed by LLM

#100

Post by d0mine »

Ukv wrote: post #87
d0mine wrote: post #78

Earlier quoted context omitted.

The original Turing game is whether machine can pretend to be a woman better than a man can (via teletype) as judged by an interrogator: > We now ask the question, "What will happen when a machine takes the part of A in this game?" Will the interrogator decide wrongly as often when the game is played like this as he does when the game is played between a man and a woman? These questions replace our original, "Can mac…

"Man pretending to be woman vs real woman" was just an example used to introduce the question in the form of a party game between humans before moving onto the actual question of a machine pretending to be human vs a real human. At a stretch, by looking at only your quoted snippet you could read that the machine is pretending to be a woman - but that interpretation is not consistent with the rest of the paper. For in…

It is consistent: "A" is a man that pretends to be a woman (B). Machine plays "A" role. Ideal machine would be indistinguishable from a man imitating a woman (it must not be too good or too bad at it compared to a man)
Post reply on HN