Knowledge Base

The knowledge base can be imagined as an open-book exam, where relevant information is retrieved from the knowledge base to serve as context for large language models when answering questions. Although the knowledge base can store large amounts of data, it's important to consider whether the uploaded data is relevant to the questions that the AI assistant needs to answer. Even with current retrieval technology, irrelevant information may still be found and used as context for answers, resulting in inability to provide appropriate responses.

For example:

  • Product information

  • FAQ (Frequently Asked Questions)

  • Other large volumes of data

Last updated

Was this helpful?