Knowledge

Knowledge can be thought of as an open-book exam: find the information from the knowledge that is most relevant to the question and extract it to serve as the context the large language model references when answering. Although the knowledge can hold a large amount of data, you should still pay attention to whether the uploaded data is relevant to the questions the AI assistant needs to answer. If it is not relevant, given current retrieval techniques, you may still end up finding unrelated information as the answer context and therefore fail to produce an answer.

For example:

• Product Information

• FAQ Frequently Asked Questions

• Other larger volumes of data