How do you verify citations at quote level, not just document level?
PICTURE THIS: AN LLM TURN
Simple meaning
Require the model to quote spans, then check that each quote is a substring of a retrieved chunk.
WHY — Tokens instead of words?
LLMs use tokens (not full words) because it helps them:
question about RAG.
trade-offs, and what you would actually do on a GenAI / LLM project - not buzzwords.
Each piece maps to a number the network can learn.
Fixed pieces are what transformers expect as input.
STEPS — What happens step by step?
Before you speak the answer, walk the interviewer through these steps:
- 1Require the model to
quote spans, then check that each quote is a substring of a retrieved chunk.
- 2Why it exists
Reject or regenerate on mismatch.
- 3Document-level citations can be
true as IDs and still support a sentence the chunk never said.
- 4Give an example
One tiny concrete case you can say aloud.
- 5Common mistake
What juniors usually get wrong.
- 6Close
When you pick this over the alternative.
EXAMPLE — See it in action
Here's a short line you can speak, broken into clear beats:
Note: Adapt this scaffold to your own project — keep it under 60–90 seconds.
Key takeaway
Require the model to quote spans, then check that each quote is a substring of a retrieved chunk. Reject or regenerate on mismatch.