🙂: Yeah, %&#@$*!
How well does the LLM understand
the structure of our data ?
Retriever
LLM
How well does the LLM interpret user queries
in context of our data?
Database
How to retrieve all records
sufficient to reason out an accurate reponse?
How well did we align retrieved data for reasoning?
Reasoners[]
LLM
SELECT *
FROM products
WHERE category = '%&#@$*'
🔑