Transformers and Attention

Explain conceptually how attention lets a transformer weigh context and why it scales to large language models.