Comment by mike_hearn
15 hours ago
Yeah a few of them are rough matches, but I'd say their overall claim is clearly correct.
It's amusing and ironic that LLMs critique humans for copying their opinions from the crowd, while having way less diversity of answers to that question than humans would. More seriously I wonder what's going on here and would like to read some proper analysis. It's not the first time models from totally different labs, trained on separately gathered datasets, somehow converge on bizarrely similar answers to open questions. It's like another version of Elias and Elara The Lighthouse Keepers where every model wants to tell the same story with the same characters in them. The weights are initialized randomly and in theory the labs should have quite some divergence in datasets by now due to different scraping, different user feedbacks and conversations... why do they converge on nearly identical stories and opinions?
>while having way less diversity of answers to that question than humans would.
Why is this ironic? The actual number of different LLMs is tiny. And a lot of those are just distilled from SOTA models.
>The weights are initialized randomly and in theory the labs should have quite some divergence in datasets by now due to different scraping,
I think the opposite, the models actually converge over time. They ingest FAR more data than any human ever could in countless lifetimes. Any one particular bit of data may only have a part in a trillion of effect. From this data you start to reach means and statistical averages across all models.
Now, I don't always believe it will be this way. As training models get cheaper and faster we will see unique LLMs start to diverge and be used more often.
The models in agreement here are all the frontier labs. I don't believe Google is distilling Anthropic or Anthropic is distilling OpenAI.
I don't think we're seeing averages or obvious fix points in the dataset. People went looking for the origins of Elias the Lighthouse Keeper and couldn't find one. Nobody predicted this outcome, it was discovered. It's a weirdly specific character and name choice, one that isn't obviously dictated by "the training set" (whatever that is), and yet they all reached the same place. How can such an obscure point in latent space be optimal for every model??
>How can such an obscure point in latent space be optimal for every model??
While I don't have any evidence based answers here, this could be some interesting mathematical artifact of the set of human knowledge and platonic algorithms.
Kind of like how carcinization turns some animals into similar shapes and forms, maybe there are just weird algorithmic shapes that emerge from how we use language. Of course we'd need to see more of these in different languages other than English to see if they disappear in some languages or appear more often in others.
This is just another one of those things about LLMs that we have no idea why it happens, and there are so many others already. And when you look at the pace things are progressing it seems we're building up new questions like this far faster than we are getting answers for them.
Didn't Elias come from WildChat?
3 replies →
> while having way less diversity of answers to that question than humans would
Isn't that to be somewhat expected due to https://en.wikipedia.org/wiki/Aumann%27s_agreement_theorem ?
> Elias and Elara
Wow, I hadn't heard of this. Fascinating to think about parallels with humans who also seem to develop similar proto-stories in independent populations.(e.g. religions involving Invisible gods).
> e.g. religions involving Invisible gods
To be fair, it's much (much!) easier to explain away an invisible god that nobody's ever seen, than a visible god that nobody's ever seen.
Something worth repeating is that in spite of their ever-expanding competence the models remain, fundamentally, token prediction algorithms. And so given a comparable base of training material, and a comparable algorithmic implementation, you should expect to see comparable answers to most questions.
This is also why I think the 'acceleration' hype is probably unjustified. What seems vastly more likely is that we start to reach an asymptotic 'perfection' where models can accurately and consistently provide meaningful next token prediction with minimal hallucinations. And then a race downward to compact this into the most efficient package possible.
This would mean the models still would not be able to create things that do not follow in a reasonable probabilistic fashion given a knowledge of [everything], and so that is the domain, 'applied creativity', where I'd expect humans to remain dominant.
In art, history, humanities sure - you need humanities to generate data there.
But in many sciences - esp purely theoretical ones - we have infinite datasets that we can generate and there can be no ultimate convergence.
The lighthouse thing seems much more odd. That humans take social shortcuts and justify them as rational is… I mean, the only thing that makes it not the most obvious “contrarian take” about humans, is that it isn’t really all that contrarian at all.
> why do they converge on nearly identical stories and opinions?
The same reason they do when it comes to coding?
Take a big enough sample size of human opinions and there seems to be some convergence in the statistical ether.
Like, if we polled 100000 humans, what answer would we converge to?
additionally, probably strong overlap in training data