The bread-and-butter of machine learning.
This is the most vanilla neural network. Also called a fully-connected network. This is just a series of matrix multiplications with non-linearities between them.
You tweak the values in the matrices to make your loss go lower and you're learning.
A transformer is a neural network that processes a set of items. These items are typically called tokens, which comes from text tokenization. But now that Transformers are popular outside of text processing, any kind of item you process with transformers is referred to as a token.