Moderate Hybrid search Question 138 of 223

What is Maximal Marginal Relevance in retrieval?

GenAI / LLM · Speak this in 60–90 seconds · Faridabad & Delhi NCR

PICTURE THIS: AN LLM TURN

Text inTokens

TransformerAttention

Text outNext token

Simple meaning

MMR reranks hits to balance relevance with diversity so near-duplicate chunks do not fill the prompt.

WHY — Tokens instead of words?

LLMs use tokens (not full words) because it helps them:

Hybrid search questions separate

people who only read docs from people who shipped.

Keep it short, concrete,

and tied to GenAI / LLM work.

Stable token IDs

Each piece maps to a number the network can learn.

Fits the model

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:

1
MMR reranks hits to
balance relevance with diversity so near-duplicate chunks do not fill the prompt.
2
It subtracts similarity to
already selected items from the relevance score.
3
Use it when your
corpus repeats the same paragraph across many files.
4
Give an example
One tiny concrete case you can say aloud.
5
Common mistake
What juniors usually get wrong.
6
Close
When you pick this over the alternative.

EXAMPLE — See it in action

Here's a short line you can speak, broken into clear beats:

Say this line

“It subtracts similarity to already selected items from the relevance score.”

Break into beats

Itsubtractssimilaritytoalreadyselected

Speaking order

2987408337471632900

Note: Adapt this scaffold to your own project — keep it under 60–90 seconds.

Key takeaway

MMR reranks hits to balance relevance with diversity so near-duplicate chunks do not fill the prompt. It subtracts similarity to already selected items from the relevance score.