# @davidad — 2024-11-27

♥6 ↻1 · https://x.com/davidad/status/1861806222443425970

@ciphergoth yes, there are hundreds of layers between each token (*causally* between, though they are usually depicted as an orthogonal dimension of the architecture), half of which are literally called “self-attention layers” https://t.co/pNuaABv68u

tags: author:davidad, kind:tweet, thread-context, year:2024