AI-103 Practice Question 74: Fine Tuning JSONL Chat Format

Microsoft Azure AI Apps and Agents Developer Associate (AI-103) · Test 8 · question 4 of 10

Implement generative AI and agentic solutionseasy

A developer is preparing a training file to run a supervised fine-tuning job against a chat model in Microsoft Foundry. They have 200 example support exchanges and want to format the file correctly on the first attempt.

Which structure should the training file use?

Show answer and explanation

Answer:

B. A .jsonl file with one JSON object per line, each containing a messages array of system/user/assistant role objects

Microsoft's fine-tuning documentation is specific: training and validation data "must be formatted as a JSON Lines (JSONL) document in the conversational format that the Chat Completions API uses." Each line is one training example, and each example is a messages array of role-tagged turns: system for the model's behavior, user for the input, assistant for the expected response, matching exactly what a live chat completion request looks like. The file must also be UTF-8 encoded with a byte-order mark, and at least 10 examples are required to start a job.

- A is wrong because prompt/completion pairs are the older single-turn completions format, not the conversational format current chat models fine-tune on, and it isn't line-delimited. - C is wrong because CSV carries no concept of roles or multi-turn structure, and the service's training and validation files are JSONL, not tabular. - D is wrong because collapsing a conversation into one opaque text string throws away the role boundaries the service uses to know which turns are instructions, which are user input, and which are the assistant output it should learn to produce.

Reference: Customize a model with fine-tuning

Practice Test 8 ›

Study guides for AI-103

AI engineer roadmap 2026: skills, certs and the orderAI engineer roadmap for 2026: skills to learn in order, AI engineer vs ML engineer, which certifications fit each stage, and what to skip.

AI-200 exam guide: Azure AI Cloud Developer AssociateAI-200 replaced AZ-204: what the Azure AI Cloud Developer exam tests, how it differs from AI-103, price in USD and INR, format and how to prepare.

How AI Is Changing Software Jobs: Where to PrepareHow AI is changing software jobs: what history shows, the evidence on junior developers, what Klarna learned, new AI roles and five domains to prepare for.

AI-901 exam guide: Azure AI Fundamentals after AI-900AI-901 replaced AI-900: learn what changed, how much Python and Microsoft Foundry you need, exam costs in India, and how to prepare.

Best AI certifications 2026: AWS, Google, Microsoft, ClaudeThe best AI certifications in 2026 from AWS, Google, Microsoft, Claude and Databricks, by level, with price, status and which one to take first.

Microsoft AI certification path 2026: which exam firstMicrosoft AI certification path for 2026 - AI-901, AB-730, AI-103, AI-200, AI-300, AB-100, AI-500 and GH-300 by role, and which exam to take first.

AI-103 Certification Guide: What Changed From AI-102The AI-103 certification replaced AI-102 in 2026. Exam format, cost, the five skill domains, what changed and how to prepare, with free practice questions.

What Is RAG? How Retrieval-Augmented Generation WorksWhat RAG is and how it works: chunking, embeddings, vector search, hybrid search, reranking, GraphRAG and agentic RAG on AWS, Google and Azure.

Prompt engineering vs RAG vs fine-tuningWhen to use prompt engineering, retrieval-augmented generation, fine-tuning or continued pre-training, compared on cost, freshness, hallucination and privacy.