Chapter 14 Quiz12 min check

Neural Embeddings

Check whether you can choose embedding models with production judgment.

Quiz A: Concepts

1 Neural embeddings are dense vectors produced by trained ____.

2. What are neural embeddings usually strong at?

3 A vector collection must match the embedding model's dimension ____.

4. What usually happens when changing embedding models?

5 Dense retrieval is often paired with sparse retrieval in a ____ search system.

6. How should you choose an embedding model?

7 Production embedding choices must consider latency, cost, privacy, and ____.

8. Why keep sparse retrieval around?

9 Shadow testing compares the old and new indexes before full ____.

10. What should embedding traces store?

Perfect score. You are choosing embeddings like a production engineer.