Kindly remind you that LLM's father is Indian.

>>106019358 (OP)

But they gave the Nobel to someone other guy LOL

>>106019358 (OP)
Why are Indians so dysgenic?

>>106019358 (OP)
>LLM
Useless dead-end tech.

>>106019358 (OP)
>Yes saaar India #1 we invent everything
I can't wait for Indians to claim to be one of the lost tribes next.
>inb4 they already do I just haven't seen it yet

Cute. When will she start taking skittles and wearing dresses?

>>106019420
>inb4 they already do I just haven't seen it yet
https://en.wikipedia.org/wiki/P._N._Oak
They make 5%ers and Hoteps sound modest in comparison

>>106019452
>Among his prominent claims were that Christianity and Islam are both derivatives of Hinduism; that Vatican City, Kaaba, Westminster Abbey and the Taj Mahal were once Hindu temples dedicated to Shiva; and that the Papacy was originally a Vedic Priesthood.
What in the fuck did I just read?

>>106019466
i dont know but hinfuism is a fossil religion

>>106019358 (OP)
Pytorch was also created by an Indian. It's over for oinks.

Does he have a daughter ?

>>106019695
Artificial intelligence may be too difficult a concept for low-IQ pigskins to understand.

>>106019358 (OP)
>Indians are born from Indians
That's crazy.

>>106019772
>indians 112 IQ average
LOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOL

>>106019772
>>106019695
IQ is racist pseudoscience
>>>/pol/

>>106019405
You are what you eat and they literally eat shit.

>>106019358 (OP)
transformer was the real tilting point

>>106019358 (OP)
>LLM's father is Indian.
you are retarded and IA alliterated if you think that all LLMs are transformer based.

>>106019358 (OP)
You mean the transformer model and it wasn't just him, it was a larger team.

>>106019358 (OP)
Didn't know Markov is an indian name

>>106023811
>Markov is an indian name
you are severely rusty if you think modern language models are a simple Markov process, or you simply are repeating it as a parrot.
there is no way in hell that few step backs can model what a current transformer model can represent which uses the WHOLE context.

>>106023916
>you are severely rusty if you think modern multiplcation models are a simple addtion process, or you simply are repeating it as a parrot.
>there is no way in hell that few additions can model what a current multiplication model can represent which uses the WHOLE number.

>>106023985
transformer based models aren't markov process
https://safjan.com/understanding-differences-gpt-transformers-markov-models/
is not the same
The generalization of markov is not even useful to replicate the model
your analogy to additions and multiplications is totally unvalid here.

>>106023985
Transformer models are not Markov processes, and this is actually a crucial distinction that highlights one of the key innovations of the transformer architecture.
In a Markov process, the next state depends only on the current state - there's no "memory" of the full history. For language modeling, this would mean predicting the next token based only on the immediately preceding token, ignoring all earlier context.
Transformer models, however, use self-attention mechanisms that allow each position to attend to all previous positions in the sequence simultaneously. This means when predicting the next token, the model has access to the entire preceding context, not just the most recent token or state.
This non-Markovian property is what enables transformers to:

Maintain long-range dependencies across sequences
Resolve references and pronouns that might be separated by many tokens
Understand complex syntactic structures that span multiple clauses
Generate coherent text that maintains thematic consistency over long passages

The self-attention mechanism essentially creates a representation where each token's processing is informed by the entire context history, violating the Markov assumption of conditional independence given the current state.
This distinction is fundamental to understanding why transformers have been so successful compared to earlier sequential models like basic RNNs, which were more Markovian in nature (though even RNNs with hidden states technically weren't pure Markov processes either).

>>106024052
>>106024070
Thanks ChatGPT but I'm not a mathmatician and when I say Transformers are like Markov chains (not processes) I don't mean they share the retardedly precisely defined "Markov property" but that they are probabilistic models with big multidimensional matrixes as inputs.

Also as far as I can tell, from what you're saying Transformers are just optimized Markov processes (added memory feature)

>>106024124
>Transformers are just optimized Markov processes (added memory feature)
they aren't markov models
if that were the case you could model them as markov process and is not the case.

>but that they are probabilistic models with big multidimensional matrixes as inputs.
the fact that they are probabilistic models doesn't mean they are tractable under the classic probability frameworks, if you try to do the same using pure probability distributions you will miserably fail.
You use classical probability and statistics when you really understand which are the relations between your variables and know the probability distribution that may fit your data. When you have no idea which the probability distribution is , then you use deep learning , with the cost of not even knowing the probability functions.
And don't say like "muh is just gaussean mixtures" because that's another bullshit, internally the model could be fitting new kinds of distributions you don't even know

>>106019772
The smartest Europeans migrated hundreds of years ago.

>>106024531
It's not really an advantage when the model is so complex that no human knows how it works, especially when, like all current LLMs, they often produce wrong outputs.