Rendered at 11:00:58 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
QubridAI 55 minutes ago [-]
It's an interesting connection to the GPU-autoresearch post; once agents have the real infrastructure, sandboxing isn't just optional anymore it becomes a bottleneck.
NooneAtAll3 54 minutes ago [-]
I thought "data efficiency" meant same quality with less parameters
instead it's more parameters with less training data... but I don't really see any quality control?
pastescreenshot 6 hours ago [-]
The result is interesting, but the practical question for me is where the compute bill lands once you include both training and serving. If a fixed-data regime pushes you toward ensembles plus chain distillation, is the endgame “serve the ensemble”, or do you expect most of the gain can be compressed back into a single deployable model later? That seems like the difference between a neat scaling result and a generally usable recipe.
sdpmas 6 hours ago [-]
oh ensemble can be distilled to a single model easily.
SknCode 3 hours ago [-]
How?
sigmoid10 47 seconds ago [-]
Same way you distill any model. Data efficiency matters only while you train the source model.
phr4ts 1 hours ago [-]
The brain does optimization during sleep. Is that something llms can benefit from?
nsnzjznzbx 14 hours ago [-]
We will get to the point where you can quickly bootstrap i.e. an LLM can train a better LLM in a loop, leave it and it can really learn. Like learn learn.
"Train yourself to solve this problem see OBJECTIVE.md"
nine_k 12 hours ago [-]
This is the kind of runaway self-improving development that proponents of the singularity keep talking about.
The problem is that training appears to be really slow and expensive. Some quality thinking is required to improve the training approach and the architecture before committing resources to training a new large model. And even the largest models are by now not nearly as good at quality thinking as the best humans.
8 hours ago [-]
andai 12 hours ago [-]
What's the human baseline? How many cats does a human need to see to learn what a cat is, vs an AI?
Maybe not quite a fair comparison since my human brain has been "learning" for half a billion years before I was born.
I wonder if there's an equivalent of that for AI. Evolving the architectures?
ainch 10 hours ago [-]
The human genome contains around 1.5GB of information and DeepSeek v3 weighs in at around 800GB, so it's a bit apples-to-oranges. As you say, what's been evolved over hundreds of millions of years is the learning apparatus and architecture, but we largely learn online from there (with some built-in behaviours like reflexes). It's a testament to the robustness of our brains that the overwhelming majority of humans learn pretty effectively. I suspect LLM training runs are substantially more volatile (as well as suffering from the obvious data efficiency issues).
If you'd like an unsolicited recommendation, 'A Brief History of Intelligence' by Max Bennett is a good, accessible book on this topic. It explicitly draws parallels between the brain's evolution and modern AI.
nxpnsv 4 hours ago [-]
The comparison is weird as we don't think with the Genome. There are something like ~100 billion neurons with ~100 trillion connections in an adult human brain . I don't know how many bytes of sourcecode deepseek has, but I don't think it helps in determining the amount of reasoning it can do.
jack_pp 8 hours ago [-]
And that same information contained in an LLM is a compression of how many terabytes of training data? Maybe in the future there will be models an order of magnitude smaller and still better performing.
What I'm saying is you can't judge the data in the genome by purely counting the bytes of data.
idiotsecant 6 hours ago [-]
The human genome isn't its own thing, the genome as a static sequence is really just an abstraction. What actually functions as the heritable unit includes epigenetic marks, non-coding RNA regulation, 3D chromatin structure, and mitochondrial DNA. In the real biological world there are very few sharp edges - systems bleed into one another and trying to define something like 'the number of bits in the human genome' is very difficult, but it's undoubtedly way bigger than you posit here.
jamilton 8 hours ago [-]
Also interesting to consider how much "compute" has to be spent by humans are learning something like that. Like, do we need to see more examples if learning from pictures of cats and dogs than seeing them in person? How many more examples? What if we're seeing them all in sequence, or spread out across hours or days?
I've probably seen... at least a dozen pictures of aardvarks and anteaters and maybe even see one of them at the zoo but I don't think I could reliably remember which was which without a reminder.
pmontra 5 hours ago [-]
If you see one picture of a zebra, fly to Africa, see a real zebra, you recognize it as a zebra. But zebras are really unmistakable.
If you see a picture of an oryx and a picture of a kudu, maybe you remember the shape of their horns and a picture is enough.
Enter waterbucks and steenboks. That starts to require a little more training.
Go all the way from mammals to insects. Bees and wasps and ants are still in the one picture is enough category. But what species of ants those on the wall of my house belong to?
I believe that ease of detection depends on how much things stand out on their own. Anyway, we do use a fundamentally different way of training than neural nets because we don't rebuild ourselves from scratch. However birds and planes fly in totally different ways but both fly. Their ways of flying are appropriate for different tasks, reach a branch or carry people to Africa to look at zebras.
sdpmas 11 hours ago [-]
i think evolution meta-learns the architecture, hyperparams. some domain knowledge too (for ex, we all perceive the world as 3d) but not much. if you compare the text consumed by human vs AI (and i think this is fair b/c even with evolution text is a pretty recent invention for humans), the gap is many orders of magnitude.
I think my toddler saw roughly 100 dogs and cats before she was able to reliably tell the difference.
pmontra 5 hours ago [-]
That happened at toddler stage of brain development and of knowledge buildup.
Let's suppose that you meet adults that never saw cats and dogs. You show them a picture a cat and a dog. Do you expect that they need to see 100 of them before telling the difference?
abeppu 12 hours ago [-]
In their little algorithm box on Chain Distillation, they have at step 2b some expression that involves multiplying and dividing by `T`, and then they say "where α = 0.5, T = 1.0".
I think someone during the copy-editing process told them this needed to look more complicated?
arjie 9 hours ago [-]
tl;dr it makes sense once you see there are hidden softmax in there; it's just the explicit formula written out and then applied with the common param value
So CE is cross-entropy and KL is Kullback-Leibler, but then division by T is kind of silly there since it falls out of the KL formula. So considering the subject, this is probably the conversion from logits to probabilities as in Hinton's paper https://arxiv.org/pdf/1503.02531
But that means there's a hidden softmax there not specified. Very terse, if so. And then the multiplication makes sense because he says:
> Since the magnitudes of the gradients produced by the soft targets scale as 1/T2 it is important to multiply them by T2 when using both hard and soft targets.
I guess to someone familiar with the field they obviously insert the softmax there and the division by T goes inside it but boy is it confusing if you're not familiar (and I am not familiar). Particularly because they're being so explicit about writing out the full loss formula just to set T to 1 in the end. That's all consistent. In writing out the formula for probabilities q_i from logits M_k(x)_i:
q_i = exp(M_k(x)_i / T) / sum_j exp(M_k(x)_j / T)
Hinton says
> where T is a temperature that is normally set to 1. Using a higher value for T produces a softer
probability distribution over classes.
And then they're using the usual form of setting T to 1. The reason they specify the full thing is just because that's the standard loss function, and it must be the case that people in this field frequently assume softmaxes where necessary to turn logits into probabilities. In this field this must be such a common operation that writing it out just hurts readability. I would guess one of them reading this would be like "yeah, obviously you softmax, you can't KL a vector of logits".
Good question. I just sort of skipped over that when reading but what you said made me think about it.
sdpmas 12 hours ago [-]
the T stands for tea :)
naruhodo 11 hours ago [-]
Ah, so it's a source of randomness! Presumably 1.0 corresponds to a really hot cup of fresh tea.
naasking 9 hours ago [-]
Great project. On the matter of data efficiency and regularization, I'd love to see someone try scaling GrokAlign!
littlestymaar 15 hours ago [-]
> Data efficiency matters because compute grows much faster than data
[2] (referencing a paper from 2022)
I'm not convinced this is particularly true in today's world, if you have more compute, you can simply generate more, and higher quality, artificial data. That's what all labs have been doing since at least 2023.
Also, the post references the Chinchilla-optimal training as a comparison baseline, but everyone has moved far beyond Chinchilla scaling, small models are routinely trained on 10-400 times more data than (1-40T tokens) than the Chinchilla-optimal number, so the entire industry went the complete opposite of what they are proposing.
That doesn't mean the techniques presented here are useless or anything (I'm not qualified to judge) but you should take the introduction with a grain of salt.
ACCount37 13 hours ago [-]
There's "cheap" bulk data - simple synthetics, unfiltered scrapes. Used for pre-training, especially early pre-training. And then there's "expensive" data. Human domain expert solutions, made by people you hire for $100 an hour. Used for SFT.
For "expensive" data, it makes a lot of sense to use every trick in the book to squeeze that data for all its worth.
akshayvegesna 15 hours ago [-]
You seem to be making two points:
- synthetic data is a valuable direction to pursue when you have compute
- chinchilla scaling laws have some flaws for small models
Both of these are side points to the core purpose of the Slowrun.
The main point is the 100M tokens we train on push people to come up with novel ideas to improve pretraining, outside of facile synthetic data generation. I think we should continue to push on synthetic data, but why not come up with some new ideas too? You cannot use synthetic data for everything (see sdpmas's point)
sdpmas 15 hours ago [-]
> you can simply generate more, and higher quality, artificial data
this is simply not true. and it's very clear if you look at continual learning, robotics, biology, etc. each has enough economic incentives to spend 1000x compute if that led to much better results, but we just don't know how to do that.
good point on chinchilla, but our models are still absurdly large no matter what standards you compare them to.
littlestymaar 15 hours ago [-]
> this is simply not true. and it's very clear if you look at continual learning, robotics, biology, etc. each has enough economic incentives to spend 1000x compute if that led to much better results, but we just don't know how to do that
I'm (and so is the post itself) talking about LLMs in particular, and this is indeed true for LLM.
sdpmas 15 hours ago [-]
continual learning is LLMs :)
ultimately everything will be/already is data bottlenecked.
instead it's more parameters with less training data... but I don't really see any quality control?
"Train yourself to solve this problem see OBJECTIVE.md"
The problem is that training appears to be really slow and expensive. Some quality thinking is required to improve the training approach and the architecture before committing resources to training a new large model. And even the largest models are by now not nearly as good at quality thinking as the best humans.
Maybe not quite a fair comparison since my human brain has been "learning" for half a billion years before I was born.
I wonder if there's an equivalent of that for AI. Evolving the architectures?
If you'd like an unsolicited recommendation, 'A Brief History of Intelligence' by Max Bennett is a good, accessible book on this topic. It explicitly draws parallels between the brain's evolution and modern AI.
What I'm saying is you can't judge the data in the genome by purely counting the bytes of data.
I've probably seen... at least a dozen pictures of aardvarks and anteaters and maybe even see one of them at the zoo but I don't think I could reliably remember which was which without a reminder.
If you see a picture of an oryx and a picture of a kudu, maybe you remember the shape of their horns and a picture is enough.
Enter waterbucks and steenboks. That starts to require a little more training.
Go all the way from mammals to insects. Bees and wasps and ants are still in the one picture is enough category. But what species of ants those on the wall of my house belong to?
I believe that ease of detection depends on how much things stand out on their own. Anyway, we do use a fundamentally different way of training than neural nets because we don't rebuild ourselves from scratch. However birds and planes fly in totally different ways but both fly. Their ways of flying are appropriate for different tasks, reach a branch or carry people to Africa to look at zebras.
Let's suppose that you meet adults that never saw cats and dogs. You show them a picture a cat and a dog. Do you expect that they need to see 100 of them before telling the difference?
I think someone during the copy-editing process told them this needed to look more complicated?
Bloody hell, I am so unfamiliar with ML notation:
So CE is cross-entropy and KL is Kullback-Leibler, but then division by T is kind of silly there since it falls out of the KL formula. So considering the subject, this is probably the conversion from logits to probabilities as in Hinton's paper https://arxiv.org/pdf/1503.02531But that means there's a hidden softmax there not specified. Very terse, if so. And then the multiplication makes sense because he says:
> Since the magnitudes of the gradients produced by the soft targets scale as 1/T2 it is important to multiply them by T2 when using both hard and soft targets.
I guess to someone familiar with the field they obviously insert the softmax there and the division by T goes inside it but boy is it confusing if you're not familiar (and I am not familiar). Particularly because they're being so explicit about writing out the full loss formula just to set T to 1 in the end. That's all consistent. In writing out the formula for probabilities q_i from logits M_k(x)_i:
Hinton says> where T is a temperature that is normally set to 1. Using a higher value for T produces a softer probability distribution over classes.
So the real formula is
And then they're using the usual form of setting T to 1. The reason they specify the full thing is just because that's the standard loss function, and it must be the case that people in this field frequently assume softmaxes where necessary to turn logits into probabilities. In this field this must be such a common operation that writing it out just hurts readability. I would guess one of them reading this would be like "yeah, obviously you softmax, you can't KL a vector of logits".Good question. I just sort of skipped over that when reading but what you said made me think about it.
I'm not convinced this is particularly true in today's world, if you have more compute, you can simply generate more, and higher quality, artificial data. That's what all labs have been doing since at least 2023.
Also, the post references the Chinchilla-optimal training as a comparison baseline, but everyone has moved far beyond Chinchilla scaling, small models are routinely trained on 10-400 times more data than (1-40T tokens) than the Chinchilla-optimal number, so the entire industry went the complete opposite of what they are proposing.
That doesn't mean the techniques presented here are useless or anything (I'm not qualified to judge) but you should take the introduction with a grain of salt.
For "expensive" data, it makes a lot of sense to use every trick in the book to squeeze that data for all its worth.
The main point is the 100M tokens we train on push people to come up with novel ideas to improve pretraining, outside of facile synthetic data generation. I think we should continue to push on synthetic data, but why not come up with some new ideas too? You cannot use synthetic data for everything (see sdpmas's point)
this is simply not true. and it's very clear if you look at continual learning, robotics, biology, etc. each has enough economic incentives to spend 1000x compute if that led to much better results, but we just don't know how to do that.
good point on chinchilla, but our models are still absurdly large no matter what standards you compare them to.
I'm (and so is the post itself) talking about LLMs in particular, and this is indeed true for LLM.