LLMs, Tokens and Embeddings Practice Test

LLMs, Tokens and Embeddings Practice Test: all 15 questions with answers

Try each question first, then open “Show answer” to check the correct option and explanation.

1. Q1.What is a large language model mainly trained to do?

   • A.Search the internet in real time
   • B.Predict the next token given the tokens before it
   • C.Store a copy of every website
   • D.Run SQL queries
   Show answer

   Correct answer: B. Predict the next token given the tokens before it

   LLMs are pre-trained on next-token prediction over very large text datasets.

2. Q2.What is a token?

   • A.Always exactly one word
   • B.A chunk of text, such as a word or part of a word, that the model processes
   • C.A password for the API
   • D.A single character only
   Show answer

   Correct answer: B. A chunk of text, such as a word or part of a word, that the model processes

   Tokenisers split text into sub-word pieces. One English word is often about one to two tokens.

3. Q3.What is a model's context window?

   • A.The browser window where you chat
   • B.The maximum number of tokens the model can consider at once, including input and output
   • C.The time limit for a request
   • D.The number of users at once
   Show answer

   Correct answer: B. The maximum number of tokens the model can consider at once, including input and output

   Anything beyond the context window cannot be seen by the model in that request.

4. Q4.Which mechanism is at the heart of the transformer architecture?

   • A.Self-attention
   • B.Decision trees
   • C.K-means clustering
   • D.Linear regression
   Show answer

   Correct answer: A. Self-attention

   Self-attention lets each token weigh how relevant every other token is to it.

5. Q5.What is an embedding?

   • A.A compressed image file
   • B.A list of numbers (vector) that represents meaning, so similar texts end up close together
   • C.A type of prompt
   • D.A database table
   Show answer

   Correct answer: B. A list of numbers (vector) that represents meaning, so similar texts end up close together

   Embeddings map text into vector space so meaning can be compared mathematically.

6. Q6.Which measure is commonly used to compare two embeddings?

   • A.Cosine similarity
   • B.File size
   • C.Word count
   • D.Alphabetical order
   Show answer

   Correct answer: A. Cosine similarity

   Cosine similarity measures the angle between vectors: closer to 1 means more similar meaning.

7. Q7.What is fine-tuning?

   • A.Writing a longer prompt
   • B.Further training a pre-trained model on task-specific data to adapt its behaviour
   • C.Lowering the temperature
   • D.Deleting part of the model
   Show answer

   Correct answer: B. Further training a pre-trained model on task-specific data to adapt its behaviour

   Fine-tuning adjusts the model's weights using a smaller, focused dataset.

8. Q8.How does pre-training differ from fine-tuning?

   • A.They are the same thing
   • B.Pre-training uses huge general datasets; fine-tuning uses smaller task-specific data
   • C.Fine-tuning always uses more data
   • D.Pre-training happens after deployment
   Show answer

   Correct answer: B. Pre-training uses huge general datasets; fine-tuning uses smaller task-specific data

   Pre-training builds general language ability; fine-tuning specialises it.

9. Q9.Why might a model not know about a recent event?

   • A.The model is broken
   • B.Its training data ends at a knowledge cutoff date
   • C.The event was too short
   • D.Models never learn facts
   Show answer

   Correct answer: B. Its training data ends at a knowledge cutoff date

   Models only know what was in their training data unless you supply newer information in the prompt or via tools.

10. Q10.What is RLHF used for?

    • A.Compressing the model
    • B.Aligning model outputs with human preferences using human feedback
    • C.Translating code
    • D.Speeding up tokenisation
    Show answer

    Correct answer: B. Aligning model outputs with human preferences using human feedback

    Reinforcement learning from human feedback trains the model toward responses people rate as more helpful and safe.

11. Q11.Setting temperature to 0 usually makes output:

    • A.More creative
    • B.More deterministic and focused on the most likely tokens
    • C.Longer
    • D.Multilingual
    Show answer

    Correct answer: B. More deterministic and focused on the most likely tokens

    Low temperature reduces randomness, which suits extraction, classification and other precise tasks.

12. Q12.Why does API cost usually depend on tokens?

    • A.Providers charge per input and output token processed
    • B.Tokens are physical items
    • C.Cost depends only on the time of day
    • D.Tokens are free
    Show answer

    Correct answer: A. Providers charge per input and output token processed

    Most LLM APIs price usage by input and output tokens, so long prompts and long answers cost more.

13. Q13.What is a multimodal model?

    • A.A model that only handles text
    • B.A model that can process more than one type of input, such as text and images
    • C.Several copies of the same model
    • D.A model with several prompts
    Show answer

    Correct answer: B. A model that can process more than one type of input, such as text and images

    Multimodal models accept and reason over different data types, for example a photo plus a question.

14. Q14.What does the max output tokens setting control?

    • A.The size of the training data
    • B.The maximum length of the generated response
    • C.The number of users
    • D.The temperature
    Show answer

    Correct answer: B. The maximum length of the generated response

    If the limit is too low, the answer is cut off mid-way.

15. Q15.Why do chat apps stream responses token by token?

    • A.It reduces the total cost to zero
    • B.Users see text sooner, which improves perceived speed
    • C.It is the only way models work
    • D.It makes answers more accurate
    Show answer

    Correct answer: B. Users see text sooner, which improves perceived speed

    Streaming lowers time-to-first-token for the user. It does not change the answer itself.