Skip to index

GLOSSARY

Transformer

The neural-network architecture behind virtually every modern AI model — from chatbots to image generators — built around a mechanism called attention.

Introduced in the 2017 paper “Attention Is All You Need”, the transformer replaced step-by-step recurrence with attention: every part of the input can directly influence every other part, so the model captures long-range relationships and trains in parallel across thousands of GPUs. That parallelism is what made scaling to internet-sized datasets possible at all.

Everything in this directory traces back to it: GPT-style language models are transformer decoders, image generators like Stable Diffusion embed transformer blocks, and Whisper-style audio models are transformers too. When researchers talk about “scaling laws”, they mean how transformer quality predictably improves with more data, parameters and compute.

Related terms

Tools that use this

Related categories