What is Fine-Tuning a Language Model?
What is 'fine-tuning' a language model and when should you use it instead of prompt engineering?
Quick Answer
The answer is further training a model on domain-specific data to change its behaviour permanently for a task. Fine-tuning takes a pre-trained language model and continues its training on a specialized dataset, adjusting the model’s weights so it becomes permanently adapted to a particular domain or function, unlike prompt engineering which only guides the model’s output temporarily through input instructions. On the Microsoft Azure AI Fundamentals AI-900 exam, this concept tests your understanding of when to retrain a model versus relying on prompts—a common trap is assuming prompt engineering can handle all specialized tasks, but fine-tuning is required for consistent, high-stakes outputs like classifying medical records or generating legal documents. A helpful memory tip: think of fine-tuning as giving the model a permanent new skill, while prompt engineering is like giving it a temporary reminder.
⚠ Common exam trap
Candidates often confuse fine-tuning with other model customization techniques like prompt engineering or hyperparameter tuning, but the key distinction is that fine-tuning permanently alters the model's weights through additional training, whereas prompt engineering only changes the input instructions.
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
✓
Further training a model on domain-specific data to change its behaviour permanently for a task
Fine-tuning is the process of taking a pre-trained language model and further training it on a domain-specific dataset to adapt its behavior permanently for a particular task. This is used instead of prompt engineering when the task requires consistent, specialized outputs that cannot be reliably achieved through prompt instructions alone, such as classifying medical records or generating legal documents.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Fine-tuning repairs errors in a model's base training data
Why it's wrong here
Fine-tuning adjusts a pre-trained model's weights using task-specific examples to shape its behaviour and output style; it does not repair errors in the base training data, which is neither feasible nor its purpose. It is tempting because both involve training data. Fine-tuning suits consistent formatting or domain tone, whereas prompt engineering suits quick, low-cost behavioural adjustments.
- ✓
Further training a model on domain-specific data to change its behaviour permanently for a task
Why this is correct
Fine-tuning performs additional training on domain-specific data, permanently altering the model's weights and behaviour for a task. Use it when prompt engineering cannot achieve the required consistency or specialisation, not for simple formatting or tone adjustments.
- ✗
Adjusting the model's temperature setting to produce more consistent outputs
Why it's wrong here
Temperature is an inference-time sampling parameter controlling randomness; it alters no weights, so it cannot teach the model new task behaviour. It tempts because lowering temperature does yield more consistent outputs, which is the correct fix when a deployed model already performs the task but answers vary too much.
- ✗
Selecting which pre-trained model from the Azure model catalogue best suits your task
Why it's wrong here
Choosing a pre-trained model from the catalogue is model selection, not fine-tuning, which retrains an existing model's weights on your own labelled data. It tempts because catalogue selection is a genuine step when starting a project and no suitable base model exists, but it changes no parameters and adds no task-specific behaviour.
Go deeper
Related to this question
Learn chapter
Prompt Engineering Fundamentals
Key term
Prompt engineering
Prompt engineering is the practice of designing and refining input queries to AI models to get the most accurate, relevant, and useful outputs.
Key term
Model
In IT and AI, a model is a trained mathematical representation that learns patterns from data to make predictions or decisions.
About these practice questions
This AI-900 question is part of Courseiva's 985-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam dumps. Learn why practice questions differ from exam dumps →
Same concept, more angles
1 more way this is tested on AI-900
These questions test the same concept from different angles. Work through them to make sure you can recognise it however the exam phrases it.
Variation 1. A marketing team wants to use a generative AI model to produce social media posts that match their brand's specific tone and style. They have a small set of example posts written by their copywriters. Which approach should they use to customize the model's outputs without retraining the entire model?
medium- ✓ A.Prompt engineering with carefully designed instructions
- B.Fine-tuning the model on the example posts
- C.Grounding the model with a knowledge base of brand guidelines
- D.Implementing a content filter to enforce brand rules
Why A: Prompt engineering, especially few-shot prompting, allows the model to match a desired tone and style by providing example posts in the instruction. It does not involve any training or weight updates, satisfying the 'without retraining' requirement. Fine-tuning (B) requires a large dataset and involves further training, which conflicts with the constraint. Grounding (C) adds context but does not effectively customize style, and content filtering (D) is for safety, not style.
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
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.