Moderate Embeddings Question 75 of 223

What is the difference between sparse and dense vectors in retrieval?

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

PICTURE THIS: RAG CHATBOT

QuestionEmbed query

SearchCompany docs

LLMAnswer with sources

Simple meaning

Dense vectors are short learned embeddings where every dimension is used.

WHY — Tokens instead of words?

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

They want a clean

contrast on Embeddings, not two memorised paragraphs.

Say what changes for

the developer, then one case where picking wrong hurts.

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 with tokens?

Before the model can read a sentence, it goes through these steps:

1
Dense vectors are short
learned embeddings where every dimension is used.
2
Sparse vectors are high-dimensional,
mostly zeros, often tied to vocabulary terms, as in BM25 or learned sparse models.
3
Hybrid systems use both
so exact terms and paraphrases can win.
4
Context mix
Attention looks at nearby tokens together.
5
Next token
The model scores what should come next.
6
Decode
IDs turn back into readable text.

EXAMPLE — See it in action

Let's see how a real sentence is tokenized (tokens may vary by model):

Input text

“Sparse vectors are high-dimensional, mostly zeros, often tied to vocabulary term”

Tokenized output

Sparsevectorsarehighdimensionalmostly

Token IDs (example)

2987408337471632900

Note: Actual tokens and IDs depend on the tokenizer (e.g., GPT, Llama, etc.).

Key takeaway

Dense vectors are short learned embeddings where every dimension is used. Sparse vectors are high-dimensional, mostly zeros, often tied to vocabulary terms, as in BM25 or learned sparse models.