Rag And Retrieval easy
Support engineers search a RAG system for error codes such as 'ERR-40213'. Dense-only retrieval keeps returning chunks about similar but different error codes. Which change best fixes this?
- A Increase the embedding dimension so the vectors can distinguish between numerically nearby codes.
- B Fine-tune the LLM so it learns every error code and can answer without any retrieval step.
- C Shrink chunks to 50 tokens so that each error code gets its own dedicated embedding vector.
- D Add lexical search such as BM25 alongside dense retrieval and merge the two result lists.
Answer: D. Add lexical search such as BM25 alongside dense retrieval and merge the two result lists.
Dense embeddings capture meaning, so 'ERR-40213' and 'ERR-40231' look nearly identical to them; **hybrid retrieval** adds exact-term matching for codes, proper nouns, and rare terms. The cost is running and fusing two indexes. Option A is tempting, but more dimensions do not make semantic vectors respect exact tokens; dense-only remains fine for purely conversational queries with no identifiers.
One system design question a day, with spaced repetition so the ones you miss come back.
Get the appRelated questions
- You must index 100M chunks using a 3,072-dimension embedding in fp32, served from RAM via HNSW. The raw vectors alone take about 1.2 TB before graph overhead. You need to cut memory several-fold while losing at most a point or two of recall@10. Which approach fits?
- A multi-tenant SaaS app stores all tenants in one vector index with a tenant_id field. Queries fetch the top 20 chunks by similarity and then drop chunks from other tenants. Small tenants often get 0-2 results. Why does this happen, and what is the fix?
- Runbooks are split into fixed 300-token chunks with no overlap. When a procedure step starts near the end of one chunk and finishes in the next, answers come back incomplete. What is the lowest-cost fix?
- Your offline eval shows recall@5 of 92% on a labeled set, so the gold passage is in the context for nearly every question. Yet human graders mark 35% of answers wrong, and the wrong answers often ignore or contradict the gold passage. Where should the team invest next?
- Your product requires every RAG answer to include clickable source citations that users can verify. How should those citations be produced?
- A runbook RAG system uses fixed 500-token chunks with no overlap. Users who ask for a multi-step recovery procedure get answers that stop at step 4 of 9, and retrieval logs show the top hit ends mid-procedure. What is the most targeted fix?