Embedding Projector is interactive content for projecting high-dimensional embedding vectors into 2D or 3D and exploring their structure as a point cloud. Using projections such as PCA and t-SNE, you can inspect neighborhoods, clusters, and outliers.
Selecting a point reveals metadata and nearby points, and the interface also supports labels, color coding, night mode, 3D labels, and range selection. It is a useful learning tool for investigating how embedding spaces organize words, images, text, and other data used by machine learning models.
