NCA-GENL Software Development Practice Question
A developer wants an LLM application to answer questions about an internal knowledge base that changes daily. Rather than retraining the model, they plan to retrieve relevant passages at query time and place them into the prompt. Which approach are they implementing?
⚠ Common exam trap
The trap here is reading "internal knowledge base" and jumping to fine-tuning, when the daily-change requirement rules out any approach that stores knowledge in model weights.
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
✓
Retrieval-augmented generation, where an embedding index supplies context for the prompt.
Answering over a frequently changing corpus without retraining is the defining use case for retrieval-augmented generation: an embedding index is refreshed as documents change, and retrieved passages are injected into the prompt at query time. Fine-tuning variants and few-shot prompting all fix knowledge at build time or prompt-authoring time.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Few-shot prompting, where several worked examples are prepended to every request.
Why it's wrong here
Few-shot prompting steers format and style using static examples embedded in the prompt. It does not fetch documents at query time, so it cannot answer questions about content the model never saw and gives no mechanism for keeping answers aligned with a changing knowledge base.
- ✓
Retrieval-augmented generation, where an embedding index supplies context for the prompt.
Why this is correct
Retrieval-augmented generation separates knowledge from model weights: documents are embedded into a vector index, the query retrieves the closest passages, and those passages are inserted into the prompt. Because the index can be refreshed whenever documents change, answers stay current without any retraining or fine-tuning cycle.
- ✗
Supervised fine-tuning, where labeled question-answer pairs update the model weights.
Why it's wrong here
Supervised fine-tuning bakes knowledge into the weights, so every knowledge-base change would require a new training run and redeployment. For a corpus that changes daily this is operationally unworkable, and it also gives no way to cite which source passage supported an answer.
- ✗
Parameter-efficient fine-tuning with LoRA adapters trained on the knowledge base.
Why it's wrong here
LoRA adapters reduce the cost of fine-tuning but still encode knowledge in parameters, so the artifact must be retrained as the corpus drifts. Daily updates would trigger continuous retraining pipelines, and the resulting model still cannot point to the specific passage it used.
About these practice questions
Courseiva writes every NCA-GENL question from scratch — 367 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 NCA-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 NCA-GENL exam.