Quiz on LLM Architecture and Hyperparameters

1. What does "LLM" stand for in the context of AI models?

a) Large Linguistic Model
b) Large Language Model
c) Linear Learning Model
d) Language Learning Machine

2. What is the primary purpose of the "temperature" parameter in LLMs?

a) To control the maximum length of generated text
b) To manage the randomness of the output
c) To select the most frequent words in the output
d) To adjust the number of tokens generated

3. What is "Top-p" sampling, also known as nucleus sampling, used for in LLMs?

a) To prioritize words with the highest probability
b) To limit the output to a fixed number of tokens
c) To sample from the smallest possible set whose cumulative probability is greater than p
d) To rank the output by frequency

4. How does the "Max Tokens" parameter affect the output of an LLM?

a) It sets the maximum number of unique words that can be generated
b) It limits the number of tokens in the output
c) It determines the diversity of word choices
d) It sets the temperature for output randomness

5. Which hyperparameter determines the number of top-ranked tokens considered during sampling?

a) Top-k
b) Temperature
c) Top-p
d) Max Tokens

6. What is a "token" in the context of LLMs?

a) A complete sentence
b) A single word
c) A chunk of text (like a word or part of a word)
d) A grammatical structure

7. In LLM architecture, what is the role of "attention mechanisms"?

a) To store previous outputs
b) To identify relationships between different words in the input
c) To convert text into numerical data
d) To generate random responses

8. Which of the following architectures is most commonly used in modern LLMs?

a) Recurrent Neural Networks (RNN)
b) Convolutional Neural Networks (CNN)
c) Transformer models
d) Support Vector Machines (SVM)

9. What is the main advantage of Transformer models compared to RNNs?

a) Faster computation and parallel processing
b) Better at handling images
c) Lower memory consumption
d) Automatic grammar correction

10. Which of the following is true about the relationship between "temperature" and randomness?

a) Lower temperature results in more creative outputs
b) Higher temperature produces more deterministic responses
c) Higher temperature increases randomness in the output
d) Temperature has no effect on randomness

—————————————————————-

Answer Key:

1. b) Large Language Model

2. b) To manage the randomness of the output

3. c) To sample from the smallest possible set whose cumulative probability is greater than p

4. b) It limits the number of tokens in the output

5. a) Top-k

6. c) A chunk of text (like a word or part of a word)

7. b) To identify relationships between different words in the input

8. c) Transformer models

9. a) Faster computation and parallel processing

10. c) Higher temperature increases randomness in the output