Transformer Architecture

Also: Transformer

The neural network architecture introduced in 2017 that underlies virtually all modern LLMs — characterized by attention mechanisms and parallel processing.