Transformer

Decode the acronym MHA in the transformer attention context.

Which structural property of attention is the single biggest reason transformers scaled where RNNs did not?

Define attention: what it does, and how it differs from convolution and recurrence.