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

Over fifty new hallucinations in ICLR 2026 submissions

gptzero.me

Post reply on HN
Print view View original

171–180 of 442 posts

Re: Over fifty new hallucinations in ICLR 2026 submissions

#171

Post by freehorse »

stdbrouw wrote: post #101
andy99 wrote: post #48

Earlier quoted context omitted.

I’ve reviewed a lot of papers, I don’t consider it the reviewers responsibility to manually verify all citations are real. If there was an unusual citation that was relied on heavily for the basis of the work, one would expect it to be checked. Things like broad prior work, you’d just assume it’s part of background. The reviewer is not a proofreader, they are checking the rigour and relevance of the work, which does…

The idea that references in a scientific paper should be plentiful but aren't really that important, is a consequence of a previous technological revolution: the internet. You'll find a lot of papers from, say, the '70s, with a grand total of maybe 10 references, all of them to crucial prior work, and if those references don't say what the author claims they should say (e.g. that the particular method that is employe…

It is not (just) consequence of the internet, the scientific production itself has grown exponentially. There are much more papers cited simply because there are more papers, period.

Re: Over fifty new hallucinations in ICLR 2026 submissions

#172

Post by bitwarrior »

saimiam wrote: post #165

Just today, I was working with ChatGPT to convert Hinduism's Mimamsa School's hermeneutic principles for interpreting the Vedas into custom instructions to prevent hallucinations. I'll share the custom instructions here to protect future scientists for shooting themselves in the foot with Gen AI. --- As an LLM, use strict factual discipline. Use external knowledge but never invent, fabricate, or hallucinate. Rules: L…

Are you sure this even works? My understanding is that hallucinations are a result of physics and the algorithms at play. The LLM always needs to guess what the next word will be. There is never a point where there is a word that is 100% likely to occur next.

The LLM doesn't know what "reliable" sources are, or "real knowledge". Everything it has is user text, there is nothing it knows that isn't user text. It doesn't know what "verified" knowledge is. It doesn't know what "fake data" is, it simply has its model.

Personally I think you're just as likely to fall victim to this. Perhaps moreso because now you're walking around thinking you have a solution to hallucinations.

Re: Over fifty new hallucinations in ICLR 2026 submissions

#173

Post by barfoure »

CapitalistCartr wrote: post #40
theoldgreybeard wrote: post #16

If a carpenter builds a crappy shelf “because” his power tools are not calibrated correctly - that’s a crappy carpenter, not a crappy tool. If a scientist uses an LLM to write a paper with fabricated citations - that’s a crappy scientist. AI is not the problem, laziness and negligence is. There needs to be serious social consequences to this kind of thing, otherwise we are tacitly endorsing it.

I'm an industrial electrician. A lot of poor electrical work is visible only to a fellow electrician, and sometimes only another industrial electrician. Bad technical work requires technical inspectors to criticize. Sometimes highly skilled ones.

I’d love to hear some examples of poor electrical work that you’ve come across that’s often missed or not seen.

Re: Over fifty new hallucinations in ICLR 2026 submissions

#174

Post by thaumasiotes »

zwnow wrote: post #110
thaumasiotes wrote: post #72

Earlier quoted context omitted.

Whether the information in the paper can be trusted is an entirely separate concern. Old Chinese mathematics texts are difficult to date because they often purport to be older than they are. But the contents are unaffected by this. There is a history-of-math problem, but there's no math problem.

Not really true nowadays. Stuff in whitepapers needs to be verifiable which is kinda difficult with hallucinations. Whether the students directly used LLMs or just read content online that was produced with them and cited after just shows how difficult these things made gathering information that's verifiable.

> Stuff in whitepapers needs to be verifiable which is kinda difficult with hallucinations.

That's... gibberish.

Anything you can do to verify a paper, you can do to verify the same paper with all citations scrubbed.

Whether the citations support the paper, or whether they exist at all, just doesn't have anything to do with what the paper says.

Re: Over fifty new hallucinations in ICLR 2026 submissions

#175

Post by adestefan »

zipy124 wrote: post #69
shusaku wrote: post #10

Earlier quoted context omitted.

20,000 submissions to a single conference? That is nuts

When academics are graded based on number of papers this is the result.

The problem isn't only papers it's that the world of academic computer science coalesced around conference submissions instead of journal submissions. This isn't new and was an issue 30 years ago when I was in grad school. It makes the work of conference organizes the little block holding up the entire system.

Re: Over fifty new hallucinations in ICLR 2026 submissions

#176

Post by ghaff »

TimedToasts wrote: post #96
ghaff wrote: post #21

Earlier quoted context omitted.

This is a particular meme that I really don't like. I've used em-dashes routinely for years. Do I need to stop using them because various people assume they're an AI flag?

No, but you should be prepared to have people suspect you are using AI to create your responses. C'est la vie. The good news is that it will rectify itself and soon the output will lack even these signals.

Well, I work for myself and people can either judge my work on its own merits or not. Don't care all that much.

Re: Over fifty new hallucinations in ICLR 2026 submissions

#177

Post by SauntSolaire »

peppersghost93 wrote: post #162
SauntSolaire wrote: post #153

Earlier quoted context omitted.

That's beside the point. While there may be many reasonable critiques of AI, none of them reduce the responsibilities of scientist.

Yeah this is a prime example of what I'm talking about. AI's produce trash and it's everyone else's problem to deal with.

Yes, it's the scientists problem to deal with it - that's the choice they made when they decided to use AI for their work. Again, this is what responsibility means.

Re: Over fifty new hallucinations in ICLR 2026 submissions

#178

Post by thfuran »

SauntSolaire wrote: post #153
peppersghost93 wrote: post #128

Earlier quoted context omitted.

It's a shame the slop generators don't ever have to take responsibility for the trash they've produced.

That's beside the point. While there may be many reasonable critiques of AI, none of them reduce the responsibilities of scientist.

>While there many reasonable critiques of AI

But you just said we weren’t supposed to criticize the purveyors of AI or the tools themselves.

Re: Over fifty new hallucinations in ICLR 2026 submissions

#179

Post by nialv7 »

theoldgreybeard wrote: post #16

If a carpenter builds a crappy shelf “because” his power tools are not calibrated correctly - that’s a crappy carpenter, not a crappy tool. If a scientist uses an LLM to write a paper with fabricated citations - that’s a crappy scientist. AI is not the problem, laziness and negligence is. There needs to be serious social consequences to this kind of thing, otherwise we are tacitly endorsing it.

Ah, the "guns don't kill people, people kill people" argument.

I mean sure, but having a tool that made fabrication so much easier has made the problem a lot worse, don't you think?

Re: Over fifty new hallucinations in ICLR 2026 submissions

#180

Post by noitpmeder »

merely-unlikely wrote: post #164
xvilka wrote: post #147

Earlier quoted context omitted.

Code correctness should be checked automatically with the CI and testsuite. New tests should be added. This is exactly what makes sure these stupid errors don't bother the reviewer. Same for the code formatting and documentation.

This discussion makes me think peer reviews need more automated tooling somewhat analogous to what software engineers have long relied on. For example, a tool could use an LLM to check that the citation actually substantiates the claim the paper says it does, or else flags the claim for review.

I'd go one further and say all published papers should come with a clear list of "claimed truths", and one is only able to cite said paper if they are linking in to an explicit truth.

Then you can build a true hierarchy of citation dependencies, checked 'statically', and have better indications of impact if a fundamental truth is disproven, ...

Post reply on HN