NCA-GENL Core Machine Learning and AI Knowledge Practice Question
A data scientist is preprocessing a text corpus to train a large language model. They want to convert each word into a dense vector representation that captures semantic relationships before feeding it into the transformer. Which technique should they use?
⚠ Common exam trap
Watch out — candidates often confuse tokenization with embedding, assuming that splitting text into tokens automatically provides semantic vectors.
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
✓
Word embeddings (e.g., Word2Vec, GloVe)
Word embeddings such as Word2Vec or GloVe convert words into dense vectors that encode semantic relationships, which is essential for language models to understand meaning. Tokenization is only a splitting step, while one-hot and TF-IDF yield sparse vectors lacking semantic depth. Thus, embeddings are the correct choice for capturing semantics before transformer processing.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
TF-IDF vectorization
Why it's wrong here
TF-IDF produces sparse vectors based on term frequency and inverse document frequency, reflecting word importance but not semantic similarity. It ignores context and word order, so it cannot provide the dense, context-aware representations needed for large language models.
- ✓
Word embeddings (e.g., Word2Vec, GloVe)
Why this is correct
Word embeddings like Word2Vec or GloVe generate dense, low-dimensional vectors where semantically similar words are close in vector space. They capture relationships such as king - man + woman ≈ queen, making them ideal for initializing the embedding layer of a transformer-based language model.
- ✗
Tokenization
Why it's wrong here
Tokenization splits text into tokens but does not produce dense vector representations. It is a necessary preprocessing step, but it operates at the string level and must be followed by an embedding layer to map tokens to vectors. Thus, it alone does not satisfy the requirement.
- ✗
One-hot encoding
Why it's wrong here
One-hot encoding creates sparse, high-dimensional vectors with no inherent semantic meaning, so it cannot capture relationships like similarity or analogy between words. It also scales poorly with vocabulary size, making it unsuitable for large language model training where dense, meaningful embeddings are required.
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.