Find answers from the community

Updated 3 months ago

hello guys i m using Mongodb to store

hello guys i m using Mongodb to store documents as chunks and faiss to do the similarity search when i query the response is on the source but the query says there is no answer for this question
L
R
9 comments
the LLM decided it couldn't get the information. Maybe the source information or query was to vague
for example in the documets that i have there is a lot articles and for example when i ask a question about a article with his number it says there is no article with this number but when i check the sources that gave the answer the document that i asked about is there @Logan M
@Logan M @dev_advocate what is the recommended value of similarity_top_k i dont know what is the problem
It very much depends on the data and query types, you can find the optimal value for your usecase via trial and error
Also for your query type (finding article with a number) vector searc doesn't do well
Try with metadata filters
@Rohan thank u very much is there any documentation about metadata filters ?
Add a reply
Sign up and join the conversation on Discord