NCA-GENL Data Analysis and Visualization Practice Question
A team is fine-tuning an LLM and wants to detect whether individual training examples are causing unusually large gradient updates. They plan to visualize per-example gradient norms alongside other diagnostics. Which two visualizations are most appropriate for identifying these influential examples? (Choose two.)
⚠ Common exam trap
The trap here is selecting aggregate training curves, which summarize overall progress but cannot identify which individual examples produce large gradient updates.
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
✓
A sorted bar chart of per-example gradient norms with examples ranked from largest to smallest
Detecting influential examples requires per-example gradient information. A sorted bar chart ranks examples so the largest norms stand out, while a scatter plot of gradient norm versus loss adds context about why those norms are large. Together they let the team isolate and inspect the examples driving unusually large updates.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✓
A sorted bar chart of per-example gradient norms with examples ranked from largest to smallest
Why this is correct
Sorting per-example gradient norms makes the most influential examples immediately visible at one end of the chart. For fine-tuning diagnostics, this ranking lets the team inspect the specific examples driving large updates, which is exactly the goal of detecting unusually influential training data.
- ✓
A scatter plot of per-example gradient norm versus training loss for each example
Why this is correct
Plotting gradient norm against loss reveals whether large updates come from high-loss examples or from other factors, adding diagnostic context that a ranking alone cannot provide. Outliers in this two-dimensional view are strong candidates for the influential examples the team wants to find.
- ✗
A line chart of the moving average of total training loss across epochs
Why it's wrong here
A moving average of total loss tracks overall optimization progress but aggregates away per-example information. It cannot identify which individual training examples produce large gradient updates, so it does not answer the team's specific diagnostic question.
- ✗
A pie chart showing the proportion of examples in each loss decile
Why it's wrong here
A pie chart of loss deciles summarizes the distribution of losses but discards gradient information entirely. Even if high-loss examples are overrepresented, the chart cannot show which examples have unusually large gradient norms, making it unsuitable for the stated task.
- ✗
A heatmap of the model's attention weights for a single randomly chosen example
Why it's wrong here
Attention weights for one random example describe internal model behavior on that example, not the distribution of gradient norms across the training set. It provides no ranking or comparison that would help the team locate unusually influential examples.
About these practice questions
This NCA-GENL question is part of Courseiva's 367-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 →
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.