Skip to main content

Table Retrieval Augmented Generation (RAG)

K
Written by Kay Chen

Table Retrieval Augmented Generation (RAG) allows Layar users to query uploaded tabular data using LLM.

For example, this table contains a description of a database schema with table names and table descriptions.

  1. Upload the table into Layar

  2. Click on the upload table to see the details

  3. Select Create Embeddings from the three-dot menu in the upper right-hand corner

  4. A progress bar will appear

  5. When the embedding creation is complete, it will say Embeddings Exist

  6. Navigate to CoAuthor

  7. Search for the table in the Sources panel and attach it to the prompt

  8. Ask a question

The LLM will use the data in the rows of the table to be able to properly extract an answer relevant to your question.

Did this answer your question?