Transformer Explainer is an interactive visualization tool for following the internal processing of a text generation model token by token. You can see how input text is tokenized, embedded, combined with position information, processed through attention and feed-forward layers, and used to predict the next token.
By looking at self-attention weights and intermediate representations, you can observe how a Transformer refers to context. It is a useful learning tool for understanding the foundation of large language models as an actual text generation process, not only through formulas or implementation details.
