Certified Natural Language Processing (NLP) Practitioner Exam
2026: Professional & Graduate-Level Practice Test — Text
Processing, Language Models, Deep Learning, Transformers, NLP
Applications & AI Ethics — 190 Multiple-Choice Questions with
Answers & Detailed Rationales
1. Question: Which of the following best describes the primary function of a language model
in NLP?
A. To translate text from one language to another.
B. To assign a probability to a sequence of words.
C. To identify the sentiment of a given document.
D. To extract named entities from a body of text.
Correct Answer: B
Explanation: A core function of a language model is to compute the probability of a
sequence of words, which is fundamental for tasks like text generation and speech
recognition.
2. Question: In the context of NLP, what is the primary goal of tokenization?
A. To convert text to lowercase.
B. To break a text into individual words or subwords.
C. To remove punctuation from a sentence.
D. To assign a part-of-speech tag to each word.
Correct Answer: B
,Explanation: Tokenization is the foundational step of segmenting a stream of text into
meaningful units, such as words, punctuation, or subword units.
3. Question: A data scientist is working with a corpus of legal documents containing many
terms like "herein," "thereof," and "whereas." Which text normalization technique would be
most appropriate to reduce vocabulary size without losing critical meaning?
A. Stemming
B. Lemmatization
C. Stop word removal
D. Lowercasing
Correct Answer: C
Explanation: While all are normalization techniques, removing high-frequency but low-
information words like these legal terms is the most direct way to reduce vocabulary.
4. Question: Which of the following is a primary advantage of using subword tokenization
(e.g., Byte-Pair Encoding) over word-level tokenization?
A. It is computationally simpler to implement.
B. It effectively handles out-of-vocabulary (OOV) words.
C. It always produces a smaller vocabulary size.
D. It preserves the exact grammatical structure of sentences.
Correct Answer: B
Explanation: Subword tokenization breaks unknown words into smaller, known subword
units, elegantly solving the OOV problem common in word-level models.
,5. Question: What is the fundamental difference between stemming and lemmatization?
A. Stemming is rule-based, while lemmatization is dictionary-based.
B. Lemmatization is faster, while stemming is more accurate.
C. Stemming works on the word level, while lemmatization works on the sentence level.
D. Lemmatization only works for English, while stemming is language-agnostic.
Correct Answer: A
Explanation: Stemming crudely chops off affixes using rules, whereas lemmatization uses a
vocabulary and morphological analysis to return a word's dictionary form (lemma).
6. Question: In a Bag-of-Words (BoW) model, what crucial information about the text is lost?
A. The vocabulary of the document.
B. The frequency of each word.
C. The order of the words.
D. The overall length of the document.
Correct Answer: C
Explanation: The BoW model represents text as an unordered collection of words, discarding
all information about word order and syntax.
7. Question: Which of the following is the primary purpose of calculating Term Frequency-
Inverse Document Frequency (TF-IDF)?
A. To measure the similarity between two documents.
, B. To weigh the importance of a word in a document relative to a corpus.
C. To classify a document into a predefined category.
D. To generate a summary of a long document.
Correct Answer: B
Explanation: TF-IDF increases the weight of words that are frequent in a document but rare
across the entire corpus, highlighting distinctive terms.
8. Question: Which vector space model represents words as dense, low-dimensional vectors
where semantically similar words are close to each other?
A. One-hot encoding
B. Bag-of-Words
C. TF-IDF
D. Word2Vec
Correct Answer: D
Explanation: Word2Vec, GloVe, and FastText are embedding techniques that learn dense
vectors, capturing semantic relationships between words.
9. Question: What does the "continuous bag-of-words" (CBOW) architecture in Word2Vec
predict?
A. The context words given the center word.
B. The center word given the context words.
C. The next word in a sequence.
D. The part-of-speech tag of the center word.
2026: Professional & Graduate-Level Practice Test — Text
Processing, Language Models, Deep Learning, Transformers, NLP
Applications & AI Ethics — 190 Multiple-Choice Questions with
Answers & Detailed Rationales
1. Question: Which of the following best describes the primary function of a language model
in NLP?
A. To translate text from one language to another.
B. To assign a probability to a sequence of words.
C. To identify the sentiment of a given document.
D. To extract named entities from a body of text.
Correct Answer: B
Explanation: A core function of a language model is to compute the probability of a
sequence of words, which is fundamental for tasks like text generation and speech
recognition.
2. Question: In the context of NLP, what is the primary goal of tokenization?
A. To convert text to lowercase.
B. To break a text into individual words or subwords.
C. To remove punctuation from a sentence.
D. To assign a part-of-speech tag to each word.
Correct Answer: B
,Explanation: Tokenization is the foundational step of segmenting a stream of text into
meaningful units, such as words, punctuation, or subword units.
3. Question: A data scientist is working with a corpus of legal documents containing many
terms like "herein," "thereof," and "whereas." Which text normalization technique would be
most appropriate to reduce vocabulary size without losing critical meaning?
A. Stemming
B. Lemmatization
C. Stop word removal
D. Lowercasing
Correct Answer: C
Explanation: While all are normalization techniques, removing high-frequency but low-
information words like these legal terms is the most direct way to reduce vocabulary.
4. Question: Which of the following is a primary advantage of using subword tokenization
(e.g., Byte-Pair Encoding) over word-level tokenization?
A. It is computationally simpler to implement.
B. It effectively handles out-of-vocabulary (OOV) words.
C. It always produces a smaller vocabulary size.
D. It preserves the exact grammatical structure of sentences.
Correct Answer: B
Explanation: Subword tokenization breaks unknown words into smaller, known subword
units, elegantly solving the OOV problem common in word-level models.
,5. Question: What is the fundamental difference between stemming and lemmatization?
A. Stemming is rule-based, while lemmatization is dictionary-based.
B. Lemmatization is faster, while stemming is more accurate.
C. Stemming works on the word level, while lemmatization works on the sentence level.
D. Lemmatization only works for English, while stemming is language-agnostic.
Correct Answer: A
Explanation: Stemming crudely chops off affixes using rules, whereas lemmatization uses a
vocabulary and morphological analysis to return a word's dictionary form (lemma).
6. Question: In a Bag-of-Words (BoW) model, what crucial information about the text is lost?
A. The vocabulary of the document.
B. The frequency of each word.
C. The order of the words.
D. The overall length of the document.
Correct Answer: C
Explanation: The BoW model represents text as an unordered collection of words, discarding
all information about word order and syntax.
7. Question: Which of the following is the primary purpose of calculating Term Frequency-
Inverse Document Frequency (TF-IDF)?
A. To measure the similarity between two documents.
, B. To weigh the importance of a word in a document relative to a corpus.
C. To classify a document into a predefined category.
D. To generate a summary of a long document.
Correct Answer: B
Explanation: TF-IDF increases the weight of words that are frequent in a document but rare
across the entire corpus, highlighting distinctive terms.
8. Question: Which vector space model represents words as dense, low-dimensional vectors
where semantically similar words are close to each other?
A. One-hot encoding
B. Bag-of-Words
C. TF-IDF
D. Word2Vec
Correct Answer: D
Explanation: Word2Vec, GloVe, and FastText are embedding techniques that learn dense
vectors, capturing semantic relationships between words.
9. Question: What does the "continuous bag-of-words" (CBOW) architecture in Word2Vec
predict?
A. The context words given the center word.
B. The center word given the context words.
C. The next word in a sequence.
D. The part-of-speech tag of the center word.
