Basic GenAI concepts · Question 11 of 12
A team switches to a new embedding model for its knowledge search but leaves the vectors already stored unchanged. What is the likely consequence?
1 answer is correct. Hard
Show answer
Answer
- A Queries and documents sit in incompatible vector spaces
Explanation
Every embedding model creates its own vector space, often even with a different dimension. Vectors from different models cannot be compared meaningfully. After switching models, all documents must be embedded again.