NCP-GENL Prompt Engineering Practice Question
A team is using an NVIDIA NeMo-based LLM to answer questions over a product manual. The model sometimes answers using general knowledge instead of the provided manual excerpts. They want to force the model to rely only on the supplied context. Which prompt engineering approach best addresses this?
⚠ Common exam trap
The trap here is believing that adding examples or tweaking sampling parameters will stop the model from using pretrained knowledge, when only an explicit grounding instruction with a refusal fallback reliably restricts it.
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
✓
Instruct the model to answer only from the provided context and to reply 'Not in the provided context' when the answer is absent.
Grounding a model in retrieved context requires an explicit instruction that restricts answers to that context and defines what to do when the answer is missing. A fallback phrase such as 'Not in the provided context' prevents the model from filling gaps with pretrained knowledge. Few-shot examples help style but do not enforce the boundary; temperature and context removal work against the goal.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✓
Instruct the model to answer only from the provided context and to reply 'Not in the provided context' when the answer is absent.
Why this is correct
An explicit grounding instruction with a fallback phrase constrains the model to the supplied excerpts and gives it a safe response when the context lacks the answer. This reduces reliance on pretrained knowledge and makes unsupported answers visible. It is the most direct prompt-level fix for context adherence in a retrieval-augmented setup.
- ✗
Increase the model's temperature so it explores more diverse answers from its pretrained knowledge.
Why it's wrong here
Raising temperature increases randomness and makes the model more likely to drift from the supplied context, not less. It does not create any mechanism to prioritize the manual excerpts. This change would worsen the problem because the model would sample more freely from its parametric knowledge rather than grounding its answer in the provided text.
- ✗
Add more few-shot examples of correct answers without changing the instruction about using the context.
Why it's wrong here
Few-shot examples can demonstrate desired answer style, but if the instruction still permits general knowledge, the model may continue to answer from pretrained memory. Examples alone do not establish a hard boundary around the provided context. Without an explicit grounding rule and fallback, the model has no signal to refuse answers that are absent from the manual excerpts.
- ✗
Shorten the prompt by removing the manual excerpts and rely on the model's product knowledge.
Why it's wrong here
Removing the excerpts eliminates the only source of authoritative product-specific information, guaranteeing that the model must use general knowledge. This directly contradicts the goal of grounding answers in the manual. It would increase hallucination risk and make answers less accurate for product-specific details that may not exist in pretrained data.
About these practice questions
Courseiva writes every NCP-GENL question from scratch — 352 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official NVIDIA exam blueprint
This NCP-GENL practice question is part of Courseiva's free NVIDIA 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 NCP-GENL exam.