hardMultiple Choice
Generative AI Leader Practice Question: A machine learning engineer is fine-tuning a…
A machine learning engineer is fine-tuning a large language model using LoRA (Low-Rank Adaptation) to reduce memory usage. During training, they notice that the model's performance on the downstream task is not improving. What is the most likely issue?
⚠ Common exam trap
Candidates often think LoRA's rank only affects memory and speed, not model capacity, leading them to overlook rank as the root cause of poor performance when the adapter is too constrained.
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
✓
The rank r of the LoRA adapter is too low to capture the task complexity
When fine-tuning with LoRA, the rank r determines the expressiveness of the low-rank adaptation matrices. If r is too low, the adapter lacks the capacity to learn the necessary task-specific features, causing the model's performance to stagnate. This is the most likely issue because the engineer observes no improvement, indicating the adapter's representational power is insufficient for the downstream task complexity.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✓
The rank r of the LoRA adapter is too low to capture the task complexity
Why this is correct
A low rank \(r\) restricts each LoRA adapter to a small low-rank update, so the weight deltas cannot represent the complexity of the downstream task. Since the stem's constraint is reduced memory usage during fine-tuning, an overly small \(r\) caps the adapter's expressive capacity, leaving training loss stagnant despite adequate data and learning rate.
- ✗
The learning rate is too high, causing the loss to oscillate
Why it's wrong here
An excessive learning rate produces oscillating or diverging loss, which is visible in the training curve, whereas the stem describes performance simply not improving. It is tempting because learning-rate tuning is the correct fix when loss values spike and fail to settle during fine-tuning.
- ✗
The LoRA adapter is applied to all layers, which slows convergence
Why it's wrong here
Applying adapters to all layers increases trainable parameters and memory, but it does not prevent learning; poorly initialised or misconfigured adapters do. It is tempting because broad layer coverage is genuinely used when a task differs substantially from the base model's pretraining distribution and needs deeper adaptation.
- ✗
The base model is too large for the dataset, causing overfitting
Why it's wrong here
Base model size does not cause overfitting; overfitting arises from too few training examples relative to trainable parameters, and LoRA freezes the base weights anyway. It is tempting because oversized models genuinely overfit when fine-tuned on tiny datasets without regularisation or adapters.
Go deeper
Related to this question
About these practice questions
This Generative AI Leader question is part of Courseiva's 1,008-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 by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This Generative AI Leader practice question is part of Courseiva's free Google Cloud 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 Generative AI Leader exam.