High Cost Question 194 of 223

How does speculative decoding reduce generation latency?

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

PICTURE THIS: A SENTENCE BECOMES TOKENS

The model does not read letters like humans. It reads these pieces, then predicts the next one.

Simple meaning

A cheap draft model proposes several tokens

1

WHY — Tokens instead of words?

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

This is a process

question about Cost.

Panels listen for order,

trade-offs, and what you would actually do on a GenAI / LLM project - not buzzwords.

Stable token IDs

Each piece maps to a number the network can learn.

Fits the model

Fixed pieces are what transformers expect as input.

2

STEPS — What happens with tokens?

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

  1. 1
    A cheap draft model

    proposes several tokens

  2. 2
    the large model verifies

    them in one parallel pass.

  3. 3
    Accepted tokens skip sequential

    large-model steps.

  4. 4
    Context mix

    Gains depend on draft quality

  5. 5
    Next token

    a bad draft wastes verification.

  6. 6
    Decode

    IDs turn back into readable text.

3

EXAMPLE — See it in action

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

Input text
“Accepted tokens skip sequential large-model steps.”
Tokenized output
Acceptedtokensskipsequentiallargemodel
Token IDs (example)
2987408337471632900

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

Key takeaway

A cheap draft model proposes several tokens the large model verifies them in one parallel pass.

Chat with us