Transformer Architecture

The neural network design underpinning nearly all modern LLMs, using stacked self-attention layers and feed-forward networks. Almost every AI coding assistant—from GitHub Copilot to Claude—is built on a transformer.