Attention lets a model weight the relevance of different tokens or patches, enabling context-aware representation and generation.
Open the related AI concept, tools, and learning paths.