Question 147 of 1,020

What Is Speculative Decoding?

This AI-900 practice question tests your understanding of describe features of generative ai workloads on azure. Read the scenario carefully and evaluate each option against the stated constraints before committing to an answer. After answering, compare your reasoning against the explanation and wrong-answer breakdown below. Once you have made your selection, read the full explanation to reinforce the concept and understand why each distractor is designed to mislead on exam day.

What is 'speculative decoding' and how does it improve LLM inference speed?

Quick Answer

The answer is speculative decoding, a technique that improves LLM inference speed by using a small, fast draft model to generate multiple candidate tokens in sequence, which a large target model then verifies in parallel. This works because the large model can accept or reject entire blocks of tokens at once, drastically reducing the number of slow, sequential autoregressive steps needed. On the Microsoft Azure AI Fundamentals AI-900 exam, this concept tests your understanding of how Azure AI services optimize throughput without sacrificing output quality—a common trap is thinking the small model replaces the large one, when in fact it only proposes candidates for verification. Remember the memory tip: “Draft then verify in parallel” to recall that the small model drafts, and the large model checks the whole block at once.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

Using a small draft model to generate candidate tokens that a large model verifies in parallel — improving throughput

Speculative decoding improves LLM inference speed by using a small, fast draft model to generate multiple candidate tokens in sequence, which are then verified in parallel by the large target model. This parallel verification allows the large model to accept or reject entire blocks of tokens at once, significantly reducing the number of sequential autoregressive steps required. The technique leverages the observation that draft models can produce acceptable continuations most of the time, and the large model only needs to correct mistakes, leading to higher throughput without sacrificing output quality.

Key principle: Answer the scenario, not the keyword: identify the specific constraint before choosing the most familiar-sounding option.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

  • Predicting user input before they finish typing to pre-compute responses

    Why it's wrong here

    Predictive input is a UI feature — speculative decoding is a model inference technique for generating multiple tokens per pass.

  • Using a small draft model to generate candidate tokens that a large model verifies in parallel — improving throughput

    Why this is correct

    Speculative decoding gets multiple tokens per main model pass — reducing latency without changing output quality.

    Related concept

    Read the scenario before looking for a memorised answer.

  • Generating speculative forecasts about future events using language model knowledge

    Why it's wrong here

    Future event prediction is a language model application — speculative decoding is an inference efficiency technique.

  • Running model inference on the CPU while the GPU processes the next request in parallel

    Why it's wrong here

    Heterogeneous compute parallelism is hardware scheduling — speculative decoding is a token generation strategy for efficiency.

Common exam traps

Common exam trap: answer the scenario, not the keyword

The trap here is that candidates confuse speculative decoding with simple input prediction or CPU/GPU offloading, but Microsoft often tests the specific mechanism of using a draft model for parallel token verification as the defining characteristic of speculative decoding.

Detailed technical explanation

How to think about this question

Under the hood, speculative decoding works by having the draft model autoregressively produce K candidate tokens (e.g., 4–8 tokens) using a fast, lightweight architecture like a smaller transformer or even an n-gram model. The large target model then processes these K tokens in a single forward pass (parallel verification) using a modified attention mask, computing logits for each position and accepting tokens that match its own distribution; rejected tokens trigger a rollback to the last accepted position. In practice, this can yield 2–3x speedups on latency-critical applications like real-time chat, where the draft model's acceptance rate is high due to the large model's tendency to agree with plausible continuations.

KKey Concepts to Remember

  • Read the scenario before looking for a memorised answer.
  • Find the constraint that changes the correct option.
  • Eliminate answers that are true in general but not in this case.

TExam Day Tips

  • Watch for words such as best, first, most likely and least administrative effort.
  • Review why wrong options are wrong, not only why the correct option is correct.

Key takeaway

Answer the scenario, not the keyword: identify the specific constraint before choosing the most familiar-sounding option.

Real-world example

How this comes up in practice

An e-commerce site experiences heavy traffic on Black Friday and near-zero traffic during off-peak weeks. Rather than provisioning permanent large VMs, the team uses auto-scaling groups that add capacity automatically under load and reduce it overnight. Questions like this test whether you understand elasticity, availability zones, and cloud compute scaling patterns.

What to study next

Got this wrong? Here's your next step.

Identify which exam domain this question belongs to, review the core concept, then practise similar questions from the same domain.

Related practice questions

Related AI-900 practice-question pages

Use these pages to review the topic behind this question. This is how one missed question becomes focused revision.

Practice this exam

Start a free AI-900 practice session

Short sessions build daily habit. Longer sessions build exam-day stamina. Try a timed session to simulate real conditions.

FAQ

Questions learners often ask

What does this AI-900 question test?

Describe features of generative AI workloads on Azure — This question tests Describe features of generative AI workloads on Azure — Read the scenario before looking for a memorised answer..

What is the correct answer to this question?

The correct answer is: Using a small draft model to generate candidate tokens that a large model verifies in parallel — improving throughput — Speculative decoding improves LLM inference speed by using a small, fast draft model to generate multiple candidate tokens in sequence, which are then verified in parallel by the large target model. This parallel verification allows the large model to accept or reject entire blocks of tokens at once, significantly reducing the number of sequential autoregressive steps required. The technique leverages the observation that draft models can produce acceptable continuations most of the time, and the large model only needs to correct mistakes, leading to higher throughput without sacrificing output quality.

What should I do if I get this AI-900 question wrong?

Identify which exam domain this question belongs to, review the core concept, then practise similar questions from the same domain.

What is the key concept behind this question?

Read the scenario before looking for a memorised answer.

About these practice questions

Courseiva creates original exam-style practice questions with explanations and wrong-answer analysis. It does not publish real exam questions, exam dumps, or protected exam content. Learn why practice questions differ from exam dumps →

How Courseiva writes practice questions · Editorial policy

Keep practising

More AI-900 practice questions

Last reviewed: Jun 30, 2026

Question Discussion

Share a tip, memory trick, or ask about the reasoning behind this question. Do not post real exam questions, leaked content, braindumps, or copyrighted exam material. Comments are moderated and may be removed without notice.

Loading comments…

Sign in to join the discussion.

This AI-900 practice question is part of Courseiva's free Microsoft certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI-900 exam.