Introduction to Large Language Models
Assignment- 1
Number of questions: 10 Total marks: 10 X 1 = 10
_________________________________________________________________________
QUESTION 1:
Based on Distributional Semantics, which of the following statements is/are true?
(i) The meaning of a word is defined by its relationship to other words.
(ii) The meaning of a word does not rely on its surrounding context.
a. Both (i) and (ii) are correct
b. Only (ii) is correct
c. Only (i) is correct
d. Neither (i) nor (ii) is correct
Correct Answer: d
Solution: According to Distributional Semantics, the meanings of words can be derived from
their distributional properties in a large corpora of text. It relies on the context in which words
appear.
_________________________________________________________________________
QUESTION 2:
Which of the following words have multiple senses?
a. light
b. order
c. letter
d. buffalo
Correct Answer: a, b, c, d
Solution: Each of these words has multiple meanings.
● Light can mean illumination, something not heavy, or a pale color.
● Order can refer to a sequence, a command, or a request for goods.
● Letter can mean an alphabet character or a written message.
● Buffalo can also have different senses. As an attributive noun (acting as an
adjective), it refers to a specific place named Buffalo, such as the city of Buffalo, New
York; as the verb to buffalo, means “to bully, harass, or intimidate" or "to baffle"; and
as a noun it refers to the animal (either the true buffalo or the bison).
_________________________________________________________________________
QUESTION 3:
Consider the following sentences:
Sentence 1: Amit forgot to set an alarm last night.
Sentence 2: Amit woke up late today.
Does Sentence 1 entail Sentence 2?
a. True
b. False
Correct Answer: b
Solution: Amit forgetting the alarm does not necessarily mean he woke up late.
_________________________________________________________________________
QUESTION 4:
What issues can be observed in the following text?
On a much-needed #workcation in beautiful Goa. Workin & chillin by d waves!
a. Idioms
b. Non-standard English
c. Tricky Entity Names
d. Neologisms
Correct Answer: b, d
Solution: We can observe usage of non-standard English such as “Workin” and “chillin”,
neologisms such as “workcation” in the text.
_________________________________________________________________________
QUESTION 5:
Consider the following sentences:
Sentence 1: The bats flew out of the cave at sunset.
Sentence 2: Rohan bought a new bat to practice cricket.
Question: Does the word "bat" have the same meaning in both sentences?
a. Yes
b. No
Correct Answer: b
Solution: In Sentence 1, "bat" refers to the flying mammal, while in Sentence 2, "bat" refers
to a piece of sports equipment used in cricket.
_________________________________________________________________________
QUESTION 6:
Which of the following statements is/are true?
a. Apple is a hypernym of fruit.
b. Leaf is a meronym of tree.
c. Flower is a holonym of petal.
d. Parrot is a hyponym of bird.
Correct Answer: b, c, d
Solution: Hyponymy (subset relation), Hypernymy (superset relation), Meronymy (part-of
relation), Holonymy (has-a relation)
_____________________________________________________________
QUESTION 7:
___________ deals with word formation and internal structure of words.
a. Pragmatics
b. Discourse
c. Semantics
d. Morphology
Correct Answer: d
Solution: Morphology is the study of words, how they are formed, and their relationship to
other words in the same language. It analyzes the structure of words and parts of words,
such as stems, root words, prefixes, and suffixes.
_____________________________________________________________
QUESTION 8:
Consider the following sentences:
Sentence 1: Priya told Meera that she had completed the report on time.
Sentence 2: Meera was impressed by her dedication.
Which of the following statements is/are true?
a. In Sentence 1, “she” refers to Meera.
b. In Sentence 1, “she” refers to Priya.
c. In Sentence 2, “her” refers to Priya.
d. In Sentence 2, “her” refers to Meera.
Correct Answer: b, c
Solution: This is an example of co-reference resolution in natural language processing. In
Sentence 1, "she" refers to Priya as per the structure of the sentence where Priya is the
subject, and the report was completed by her. "her" in the second sentence also refers to
Priya.
_____________________________________________________________
QUESTION 9:
In semantic role labelling, we determine the semantic role of each argument with respect to
the ___________ of the sentence.
a. noun phrase
b. subject
c. predicate
d. adjunct
Correct Answer: c
Solution: In semantic role labelling, we determine the semantic role of each argument with
respect to the predicate of the sentence.
_________________________________________________________________________
QUESTION 10:
Which of the following statements is/are true?
(i) Artificial Intelligence (AI) is a sub-field of Machine Learning.
(ii) LLMs are deep neural networks for processing text.
(iii) Generative AI (GenAI) involves only Large Language Models (LLMs).
a. Only (i) and (ii) are correct
b. Only (ii) is correct
c. Only (ii) and (iii) are correct
d. All of (i), (ii), and (iii) are correct
e. Neither (i), (ii), or (iii) is correct
Correct Answer: b
Solution: Let’s analyze each statement.
(i) Artificial Intelligence (AI) is a sub-field of Machine Learning.
Incorrect: AI is a broad field encompassing various techniques for making machines perform
intelligent tasks. Machine Learning (ML) is a sub-field of AI that focuses on learning patterns
from data. Since AI includes rule-based systems, expert systems, and other non-ML
approaches, it is incorrect to say that AI is a sub-field of ML. Instead, ML is a sub-field of AI.
(ii) LLMs are deep neural networks for processing text.
Correct: Large Language Models (LLMs) are a class of deep neural networks specifically
designed for text processing tasks such as language generation, translation, summarization,
etc. "Processing" in this context refers to how LLMs transform input text into meaningful
representations before generating an output. Note that not all classes of transformer-based
language models are trained to generate text. For example, encoder-only models are
particularly useful for creating contextual encoding of sentences/sequences (e.g.,
SentenceBERT is a popularly used model for getting sentence encoding). Autoregressive
LMs are scaled to huge sizes nowadays to “generate text”, but several other text processing
tasks are also performed by LLMs, and transformer-based LMs in general.
(iii) Generative AI (GenAI) involves only Large Language Models (LLMs).
Incorrect: While LLMs are a major part of Generative AI, they are not the only models used
in this domain. Generative AI includes models for various data modalities, such as image
generation (e.g., DALL-E, Stable Diffusion), audio generation (e.g., WaveNet), and video
generation. GenAI encompasses a broader range of generative models beyond just LLMs.
_________________________________________________________________________
