Learn vocabulary for diagnosing and fixing prompts: instruction-following failures, hallucination, format errors, negative constraints, and grounding.
0 / 14 completed
1 / 14
What does 'the model is not following instructions' mean in prompt debugging?
Instruction-following failure means the model's response violates one or more of the constraints or directives in the prompt. Common fixes include making instructions more explicit, moving them to a system message, or rephrasing to reduce ambiguity.
2 / 14
What does it mean when you say 'it's hallucinating facts'?
Hallucination refers to confident generation of false or fabricated information. Mitigation strategies include grounding the prompt with retrieved facts, instructing the model to say 'I don't know' when uncertain, or using retrieval-augmented generation.
3 / 14
What is 'iterative refinement' in prompt engineering?
Iterative refinement is the core debugging loop in prompt engineering: run the prompt, evaluate the output, identify what went wrong (wrong format, hallucination, missed constraint), adjust the prompt, and repeat until the output is satisfactory.
4 / 14
What is a 'negative constraint' in a prompt?
Negative constraints explicitly forbid certain outputs or behaviours. They are often needed when the model defaults to a behaviour you want to prevent — for instance, adding unsolicited caveats, using a particular format, or speculating beyond the provided data.
5 / 14
What does 'grounding the output' mean in prompt debugging?
Grounding means giving the model concrete evidence or data to draw from, and instructing it to base its response on that material. This reduces hallucination by constraining the model to what is actually present in the provided context.
6 / 14
Reviewer: 'I'm seeing a lot of 'off-track' responses here. The model seems to be generating code snippets that aren't related to the user's request for a simple API endpoint validation function. It needs more focus.'
Which of the following best describes what Reviewer is referring to in this comment, concerning prompt debugging?
Reviewer is highlighting that the prompt isn't effectively guiding the model. 'Off-track' in this context refers to the model generating irrelevant or unrelated outputs due to a lack of specific instructions. It's not about the training data itself (option B) or fundamental architectural issues (option A), but rather the quality and precision of the prompt's guidance, which is a core concept in iterative refinement – making sure the model knows exactly what you want.
7 / 14
Reviewer: 'I'm seeing a lot of 'off-track' responses here. The model seems to be generating code snippets that aren't related to the user's request for a simple API endpoint validation function. It needs more focus.'
Which of the following best describes what Reviewer is referring to in this comment, concerning prompt debugging?
Reviewer is highlighting that the prompt isn't effectively guiding the model. 'Off-track' in this context refers to the model generating irrelevant or unrelated outputs due to a lack of specific instructions. It's not about the training data itself (option B) or fundamental architectural issues (option A), but rather the quality and precision of the prompt's guidance, which is a core concept in iterative refinement – making sure the model knows exactly what you want.
8 / 14
Reviewer: 'I'm seeing a lot of 'off-track' responses here. The model seems to be generating code snippets that aren't related to the user's request for a simple API endpoint validation function. It needs more focus.'
Which of the following best describes what Reviewer is referring to in this comment, concerning prompt debugging?
Reviewer is highlighting that the prompt isn't effectively guiding the model. 'Off-track' in this context refers to the model generating irrelevant or unrelated outputs due to a lack of specific instructions. It's not about the training data itself (option B) or fundamental architectural issues (option A), but rather the quality and precision of the prompt's guidance, which is a core concept in iterative refinement – making sure the model knows exactly what you want.
9 / 14
Reviewer: 'I'm seeing a lot of 'off-track' responses here. The model seems to be generating code snippets that aren't related to the user's request for a simple API endpoint validation function. It needs more focus.'
Which of the following best describes what Reviewer is referring to in this comment, concerning prompt debugging?
Reviewer is highlighting that the prompt isn't effectively guiding the model. 'Off-track' in this context refers to the model generating irrelevant or unrelated outputs due to a lack of specific instructions. It's not about the training data itself (option B) or fundamental architectural issues (option A), but rather the quality and precision of the prompt's guidance, which is a core concept in iterative refinement – making sure the model knows exactly what you want.
10 / 14
Alex (Lead Developer) comments on a PR draft: 'This response is completely irrelevant. The user asked for a function to check if an API endpoint exists and returns a 200 status code. It's just spitting out HTML markup!' What does 'off-track' mean in this context?
'Off-track' in prompt debugging refers to when the model's generated output doesn't address the core request. It signifies a divergence from the intended task—in this case, generating code related to API endpoint validation—and produces something irrelevant like HTML markup. Misconceptions often arise thinking it means technical issues with the API itself; it's about the *content* of the response.
11 / 14
Sarah (Junior Developer) sends a Slack message: 'I'm struggling to get the model to consistently generate valid JSON. It keeps returning strings with brackets and quotes.' What is Sarah most likely referring to when she says 'it's not returning valid JSON'?
Sarah is concerned about the *structure* and *syntax* of the model's output. 'Valid JSON' means the response conforms to the JSON specification—specifically, it has correct syntax including proper use of commas, data types (strings, numbers), and enclosing brackets/quotes. It's a common issue with large language models that can 'hallucinate' formatting details.
12 / 14
You're debugging a prompt for an API call to retrieve customer data. The model consistently returns results including personally identifiable information (PII) like addresses and phone numbers, despite your explicit negative constraint in the prompt: 'Do not include any PII.' What technique is most relevant to address this issue?
'Grounding' the output means providing the model with reliable sources of information to anchor its responses. In this case, supplying a curated dataset will help the model understand what constitutes valid customer data and prevent it from generating PII—even if it initially ignored the negative constraint. Reducing the prompt length or disabling constraints won't solve the fundamental problem.
13 / 14
Ben (Senior Developer) writes in a PR description: 'I'm adding a new negative constraint to the prompt: 'The response should *not* include any code for database queries.' What is the primary purpose of this negative constraint?
A negative constraint is a directive to *exclude* certain types of content from the model's output. In this case, it aims to prevent the generation of database query code—which might be insecure or irrelevant to the user's request. It's not about forcing functional code or improving accuracy; those are broader goals.
14 / 14
You're reviewing a model's output for a prompt requesting a Python function to calculate the factorial of a number. The output includes comments like '// This is just a placeholder.' What does this indicate about the debugging process?
The presence of placeholder comments like '// This is just a placeholder' suggests that the model hasn't fully understood or refined its response. 'Iterative refinement' involves repeatedly adjusting the prompt and evaluating the output until it meets the desired criteria—essentially removing these unnecessary elements to achieve a more focused, useful solution.
What will I practice in "Prompt Debugging — Vocabulary"?
This is an AI Prompting exercise set. It walks through 14 scenario-based multiple-choice questions built around real usage of AI Prompting terminology that IT professionals encounter on the job.
Is this exercise free to use?
Yes. Every exercise on CoderSlingo, including this one, is free to complete with no account, sign-up, or paywall.
How many questions are in this exercise?
This set contains 14 questions. Each one shows immediate feedback and a detailed explanation after you answer, so you learn the correct usage right away rather than waiting for a final score.
Do I need prior experience to complete this exercise?
No prior experience is required. Each question includes a full explanation covering the reasoning behind the correct answer, so the exercise itself teaches the AI Prompting vocabulary as you go.
Can I retry the exercise if I get questions wrong?
Yes — use the "Try again" button on the results screen to reset your answers and go through all the questions again. There is no limit on attempts.
Is my progress saved?
Your answers and score for the current session are tracked in the browser as you go. No account or login is needed, and there is nothing to install.
What if I don't understand a term used in a question?
Read the explanation shown after you answer each question — it breaks down the correct term in plain English with a real-world example. You can also check the site Glossary for quick definitions.
How is this different from reading a blog article on the topic?
Exercises like this one are interactive drills that test and reinforce specific vocabulary through multiple-choice questions, while blog articles explain concepts in prose. Practising here after reading builds active recall, not just passive recognition.
Where can I find more AI Prompting exercises?
See the AI Prompting exercises hub for the full set of related pages, or browse all exercise categories from the main Exercises index.
Can I use this exercise to prepare for a technical interview?
Yes — AI Prompting vocabulary comes up often in technical discussions and interviews. Pair this exercise with our dedicated Interview Preparation section for role-specific practice.