- Print
- DarkLight
- PDF
Intro
In Paradigm, three essential components help manage and analyze document databases within a company or workspace: Collections, Document Chunks, and Documents.
Collections
Collections lists all document databases associated with a company or workspace. Each collection has parameters that influence the processing of associated documents, including the database owner, the associated company, the embedding model used, and the vector database.
Document Chunks
Document Chunks are excerpts generated when a user asks a question using the "chat with doc" feature. Paradigm analyzes all documents and creates these excerpts to justify the model's response. This page provides detailed information about the generated excerpts, including the identified content and associated metadata.
Documents
Documents undergo processing during the embedding/parsing phase when added to a collection. For each document, you can view comprehensive details such as the document title, text content, document owner, associated company/workspace, word count, status, format, and the number of times it has been used to answer a question. This information helps understand how the model qualifies the document and provides in-depth insights into it.