Transformer

The neural network architecture that powers all modern AI — invented at Google in 2017, named for its self-attention mechanism.

Definition

The transformer architecture (from "Attention Is All You Need", Vaswani et al. 2017) underlies every major LLM. Key innovation: self-attention, which lets each token "look at" every other token. This unlocked scaling.

Examples

  • All modern LLMs and most image/audio models use transformers

Related terms

Looking for actual gigs that use Transformer? Browse all live listings or filter by $100+/hr.