Attention Is All You Need

The 2017 paper by Vaswani, Shazeer, Uszkoreit, and five other Google Brain researchers introducing the Transformer architecture — replacing recurrent networks with self-attention — which became the foundation for GPT, BERT, and all subsequent large language models.

Introduced the Transformer — a neural network architecture that processes entire sequences simultaneously through self-attention rather than sequentially through RNNs — enabling the parallelization that made training large language models on modern hardware feasible.