3 ms·
An LLM has nothing in common with a human. An LLM works in ways that have nothing in common with the human brain.
by user_named 3y ago
An LLM has nothing in common with a human. An LLM works in ways that have nothing in common with the human brain.
- eru 3y agoYes, LLMs and human brains share at most some faint similarities. Nevertheless, human feats can act as an existence proof of what is possible. Including of what might be possible for a neural network. (I'm not sure whether a large language model necessarily needs to be a neural network in the sense of a bunch of linear transformations interleaved with some simple non-linear activation functions. But for the sake of strengthening your argument, let's assume that we are assuming this restrictive definition of LLM.)