Hacker News Reader

Comment by xg15 | Hacker News Reader

What I would really like to see more in the future is LLMs trained from scratch (i.e. not finetuned) on controlled training sets.

Currently, whenever a LLM produces a result that looks like deeper reasoning, it's easy to explain that with "eh, there was probably something similar in the trainset that the LLM just had to tweak slightly" - and that argument is basically impossible to refute or confirm, because the trainsets are huge, chaotic and often only partially available (Books2...). Therefore we can't really make any good statements about what information could or could not be pulled from the trainset.

I think it would be interesting to pretrain an LLM on a selective corpus - say, the entirety of Discworld or A Song of Ice and Fire novels - and then observe if the model can produce output that is not obviously stitched together from the books.

View on HN

> eh, there was probably something similar in the trainset that the LLM just had to tweak slightly

Look for examples of mixed image-text generative AIs, and see how they are able to understand visual jokes and memes and explain the source of their humour.

To do that, the AI not only has to recognize and classify the items in the image. It also has to identify the context of the items in the image, and figure out what elements are unusual in that context.

All that process is something that could be only imitated if the training set contained exactly the same meme; as the source of humour will be different for each image. But the AI is capable of doing it even for memes not included in the training set.

There are datasets like this: https://old.reddit.com/r/ExplainTheJoke/ and books on writing jokes etc. I can see why people are skeptical, not to say it isn't curious the models can do what they do, but I understand what the parent is asking for.

> What has been will be again, what has been done will be done again; there is nothing new under the sun.

Arguably, most everything anyone can encounter has happened to someone, and likely been written about.

Imagine a toddler learning from everything that has happened everywhere it happened all at once, as recorded in a journal.

Now, by contrast, a toddler learning from Discworld. Is that remotely enough? How much is enough?

Turns out you can look this up, as it has been graphed: it's rather more than Discworld to get these "emergent" properties. Just as it would be rather more than Discworld before a toddler without our everything journal could make sense of Discworld itself.

To your point, before reading Discworld, our toddler would need a great deal of huge chaotic training sets. Stands to reason, so might the LLM.

And now, this: https://www.quantamagazine.org/some-neural-networks-learn-la...

Reply on news.ycombinator.com