What’s the best way of getting a GPT to read its knowledge inside a GPT, when responding to a user, rather than answering from its own base knowledge
#reading knowledge inside GPT
tired notch
dusk martenBOT
Hi @tired notch. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:
To get a GPT to read its knowledge base when responding to a user, you can use the knowledge feature in the GPT. This feature allows you to upload files containing additional context that the GPT can access in response to user prompts. Here's how it works:
-
Use the GPT editor to attach up to 20 files to a GPT. Each file can be up to 512 MB in size and can contain up to 2,000,000 tokens. You can include files containing images, but only the text is currently processed.
-
When you upload a file, the GPT breaks the text up into chunks, creates embeddings (a mathematical way of representing text), and stores them for later use.
-
When a user interacts with your GPT, the GPT can access the uploaded files to get additional context to augment the user’s query. The GPT chooses one of the following methods based on the requirements of the user’s prompt:
-
Semantic search - Returns relevant text chunks as described above. This is preferred when responding to “Q&A” style prompts, where a specific portion of the source document is required.
-
Document review - Entire short documents and/or relevant excerpts of larger documents are returned and included along with the prompt as additional context. This is preferred when responding to summarization or translation prompts, where the entire source document is required.
-
Currently, the only way to manage the files attached to a GPT is using the GPT Builder UI. This means it is best for applications where context changes infrequently (employee handbooks, policy documents, school curricula, etc).
Please note that there have been instances where users have reported issues with GPTs reading files from their knowledge base. Some users have found workarounds, such as making an identical copy of the knowledge base and reading from the copy instead, or changing the file format to .txt if the GPT doesn't recognize Markdown files.
For more information, you can refer to the following sources:
- Knowledge in GPTs (https://help.openai.com/en/articles/8843948-knowledge-in-gpts)
- Forum Post: GPT suddenly can't read files in its knowledge base (https://community.openai.com/t/578576)
- Forum Post: How to get a GPT to consistently reference the documents in the knowledge base? (https://community.openai.com/t/563085)