Quiz: How ChatGPT Generates an Answer

Question 1

What is the basic unit of text that a language model processes and generates?

Tokens Webpages Databases Programs

Question 2

Why is conversation context useful when interacting with ChatGPT?

It helps the system interpret information from the ongoing conversation It guarantees that every answer will be factually correct It replaces the language model It prevents the model from generating new text

Question 3

What does inference mean in the context of a trained language model?

The process of collecting training data The process of designing the ChatGPT interface The process of using a trained model to produce an output The process of permanently changing model parameters during a conversation

Question 4

What is next-token prediction?

Predicting which webpage should be displayed next Predicting which token is likely to come next Predicting what the user will ask tomorrow Predicting whether a computer will turn off

Question 5

Which statement best describes the difference between training and inference?

Training uses a trained model to answer questions, while inference teaches the model Training and inference are exactly the same process Training is how the model learns, while inference uses the trained model to produce an output Training only happens during a conversation, while inference happens before training