It literally is just statistics… wtf are you on about. It’s all just weights and matrix multiplication and tokenization
Well on one hand yes, when you’re training it your telling it to try and mimic the input as close as possible. But the result is still weights that aren’t gonna reproducte everything exactly the same as it just isn’t possible to store everything in the limited amount of entropy weights provide.
In the end, human brains aren’t that dissimilar, we also just have some weights and parameters (neurons, how sensitive they are and how many inputs they have) that then output something.
I’m not convinced that in principle this is that far from how human brains could work (they have a lot of minute differences but the end result is the same), I think that a sufficiently large, well trained and configured model would be able to work like a human brain.
Not an LLM specifically, in particular lack of backtracking and the network depth limits as well as interconnectivity limits sets hard limits on capabilities.
https://www.lesswrong.com/posts/XNBZPbxyYhmoqD87F/llms-and-computation-complexity
https://garymarcus.substack.com/p/math-is-hard-if-you-are-an-llm-and
https://arxiv.org/abs/2401.11817
Humans have a completely different memory model and a in large part a very different way of linking together learned concepts to form their world view and to develop interdisciplinary skills, allowing us to solve many kinds of highly complex tasks as long as we can keep enough of it in our memory.
It’s all just weights and matrix multiplication and tokenization
See, none of these is statistics, as such.
Weights is maybe closest but they are supposed to represent the strength of a neural connection. This is originally inspired by neurobiology.
Matrix multiplication is linear algebra and encountered in lots of contexts.
Tokenization is a thing from NLP. It’s not what one would call a statistical method.
So you can see where my advice comes from.
Certainly there is nothing here that implies any kind of averaging going on.
Why would averaging lead to repetition of stereotypes?
Anyway, it’s hard to say LLMs output what they do. GPTisms may have to do with the system prompt or they may result from the fine-tuning. Either way, they don’t seem very internet average to me.
