That was my first thought too -- instead of talk like a caveman you could turn off reasoning, with probably better results.
Additionally, LLMs do not actually operate in text; much of the thinking happens in a much higher dimensional space that just happens to be decoded as text.
So unless the LLM was trained otherwise, making it talk like a caveman is more than just theoretically turning it into a caveman.
There was a paper recently that demonstrated that you can input different human languages and the middle layers of the model end up operating on the same probabilistic vectors. It's just the encoding/decoding layers that appear to do the language management.
So the conclusion was that these middle layers have their own language and it's converting the text into this language and this decoding it. It explains why sometime the models switch to chinese when they have a lot of chinese language inputs, etc.
Pretty obvious when you think that neural networks operate with numbers and very complex formulas (by combining several simple formulas with various weights). You can map a lot of things to number (words, colors, music notes,…) but that does not means the NN is going to provide useful results.
Everything is obvious if you ignore enough of the details/problem space. I’ll read the paper rather than rely on my own thought experiments and assumptions.
Oh, Jesus Christ. I learned to write at a college with a strict style guide that taught us how to use different types of punctuation to juxtapose two ideas in one sentence. In fact, they did/do a bunch of LLM work so if anyone ever used student data to train models, I’m probably part of the reason they do that.
You sound like you’re trying to sound impressive. Like I said, I’ll read the paper.
It is text prediction. But to predict text, other things follow that need to be calculated. If you can step back just a minute, i can provide a very simple but adjacent idea that might help to intuit the complexity of “ text prediction “ .
I have a list of numbers, 0 to9, and the + , = operators. I will train my model on this dataset, except the model won’t get the list, they will get a bunch of addition problems. A lot. But every addition problem possible inside that space will not be represented, not by a long shot, and neither will every number. but still, the model will be able to solve any math problem you can form with those symbols.
It’s just predicting symbols, but to do so it had to internalize the concepts.
This gives the impression that it is doing something more than pattern matching. I think this kind of communication where some human attribute is used to name some concept in the LLM domain is causing a lot of damage, and ends up inadvertently blowing up the hype for the AI marketing...
I think what's causing a lot of damage is not attributing more of human attributes (though carefully). It's not the LLM marketing you have to worry about - that's just noise. All marketing is malicious lies and abusive bullshit, AI marketing is no different.
Care about engineering - designing and securing systems. There, the refusal to anthropomorphise LLMs is doing a lot of damage and wasted efforts, with good chunk of the industry believing in "lethal trifecta" as if it were the holy Trinity, and convinced it's something that can be solved without losing all that makes LLMs useful in the first place. A little bit of anthropomorphising LLMs, squinting your eyes and seeing them as little people on a chip, will immediately tell you these "bugs" and "vulnerabilities" are just inseparable facets of the features we care about, fundamental to general-purpose tools, and they can be mitigated and worked around (at a cost), but not solved, not any more you can solve "social engineering" or better code your employees so they're impervious to coercion or bribery, or being prompt-injected by a phone call from their loved one.
Except I actually mean to infer the concept of adding things from examples. LLMs are amply capable of applying concepts to data that matches patterns not ever expressed in the training data. It’s called inference for a reason.
Anthropomorphic descriptions are the most expressive because of the fact that LLMs based on human cultural output mimic human behaviours, intrinsically. Other terminology is not nearly as expressive when describing LLM output.
Pattern matching is the same as saying text prediction. While being technically truthy, it fails to convey the external effect. Anthropomorphic terms, while being less truthy overall, do manage to effectively convey the external effect. It does unfortunately imply an internal cause that does not follow, but the externalities are what matter in most non-philosophical contexts.
>do manage to effectively convey the external effect
But the problem is that this does not inform about the failure mode. So if I am understanding correctly, you are saying that the behavior of LLM, when it works, is like it has internalized the concepts.
But then it does not inform that it can also say stuff that completely contradicts what it said before, there by also contradicting the notion of having "internalized" the concept.
If you look at the failure modes, they very closely resemble the failure modes of humans in equivalent situations. I'd say that, in practice, anthropomorphic view is actually the most informative we have about failure modes.
You never met a person that isn’t always right or one that makes up shit to sound smart? Because that’s the pattern you are describing that is being matched.
after you go from from millions of params to billions+ models start to get weird (depending on training) just look at any number of interpretability research papers. Anthropic has some good ones.
You obviously do not speak other languages. Other cultures have different constrains and different grammar.
For example thinking in modern US English generates many thoughts, to keep correct speak at right cultural context (there is only one correct way to say People Of Color, and it changes every year, any typo makes it horribly wrong).
Some languages are far more expressive and specialized in logical conditions, conditionals, recursion and reasoning. Like eskimos have 100 words for snow, but for boolean algebra.
It is well proven that thinking in Chinese needs far less tokens!
With this caveman mod you strip out most of cultural complexities of anglosphere, make it easier for foreigners and far simpler to digest.
>Some languages are far more expressive and specialized in logical conditions, conditionals, recursion and reasoning. Like eskimos have 100 words for snow, but for boolean algebra.
Really?
Because if one accepts that computer languages are languages, then it seems that we could identify one or two that are highly specialized in logical conditions etc. Prolog springs to mind.
We have already proven that all the computing mechanism that those languages derive their semantic forms are equivalent to the Turing Machine. So C and Prolog are only different in terms of notations, not in terms of result.
Yes, really. The concept GP is alluding to is called the Sapir-Worf hypothesis, which is largely non scientific pop linguistics drivel. Elements of a much weaker version have some scientific merit.
Programming languages are not languages in the human brain nor the culture sense.
Additionally, LLMs do not actually operate in text; much of the thinking happens in a much higher dimensional space that just happens to be decoded as text.
So unless the LLM was trained otherwise, making it talk like a caveman is more than just theoretically turning it into a caveman.