AI Quietly Picks the Most Likely Word

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-explorers-foundations-AI-and-the-confidence-percent-r9a5

1. When an AI writes a sentence, how does it decide what word to put down next?

   1. It looks up the complete sentence in a giant library of existing text
   2. It asks a human for permission before writing each word
   3. It guesses which word is most likely to come next based on what it has seen before
   4. It randomly picks any word from a list

2. If you type 'I love' into an AI, why do you get very different answers each time?

   1. The AI is broken and gives random answers
   2. The AI only gives different answers on weekends
   3. Many different words can reasonably follow 'I love,' so the AI picks from several likely options
   4. The AI forgot what you typed and makes up new responses

3. What does the lesson compare probability to in order to help you understand it?

   1. A rolling dice
   2. A weather forecast
   3. A coin flip
   4. A deck of cards

4. What does 'next-token prediction' mean in simple words?

   1. Predicting which computer key will be pressed next
   2. Planning an entire story before writing it
   3. Guessing the next word that will come after the current one
   4. Remembering all the words in a sentence

5. If an AI says a word has a 90% probability of being next, what does that tell you?

   1. The word has no chance of being chosen
   2. The word is very likely to be chosen but not certain
   3. The word will definitely be chosen
   4. The word has been chosen already

6. The lesson says AI 'usually picks the highest' probability but 'sometimes it explores.' What does 'explores' mean here?

   1. The AI stops writing and takes a break
   2. The AI goes to a website to learn new words
   3. The AI searches the internet for answers
   4. The AI sometimes picks a less likely word to keep things interesting

7. A student says, 'AI must know what it wants to say before it starts writing.' Based on what you learned, is this true?

   1. Yes, AI plans out its entire response before writing any words
   2. No, AI decides each word one at a time, not knowing the full sentence ahead of time
   3. Yes, but only for short sentences
   4. No, AI never writes anything

8. What would happen if you typed 'The sky is' into an AI?

   1. The AI would guess likely words to follow, like 'blue' or 'cloudy'
   2. The AI would refuse to answer because the sentence is incomplete
   3. The AI would write a completely random sentence about anything
   4. The AI would ask you to finish the sentence

9. Why is it important that AI 'guesses' rather than 'knows' the correct answer?

   1. Guessing is faster than knowing
   2. There is no difference between guessing and knowing
   3. Guessing means AI can be wrong, but knowing means AI is always right
   4. AI doesn't truly understand—it uses patterns to make predictions, which can sometimes be wrong

10. What kind of 'model' is a language model like the one described in the lesson?

    1. A system that has learned patterns from lots of text data
    2. A person who speaks multiple languages
    3. A physical computer inside a robot
    4. A small toy that scientists play with

11. If you wanted the AI to write something very creative and surprising, what might you want it to do more often?

    1. Pick more words with lower probability (explore more)
    2. Stop writing after the first word
    3. Always pick the highest probability word
    4. Copy exactly what other websites write

12. Based on what you learned, which sentence best describes how AI generates text?

    1. AI thinks of a whole story, then types it out all at once
    2. AI writes one word, then uses that to guess the next, then uses that to guess the next, and so on
    3. AI asks a human what to write, then types it
    4. AI copies sentences from books it has memorized

13. Why might two different people get different responses from the same AI when typing the same prompt?

    1. The AI is broken for one person
    2. The AI doesn't really work
    3. The AI might choose different words from among several likely options, or its settings might vary
    4. One person is lying about what the AI said

14. What does it mean that AI writes by 'guessing the most likely next word, again and again'?

    1. AI guesses randomly and never learns
    2. AI only guesses once and then stops
    3. AI makes one big guess about the whole story
    4. AI guesses word by word, each new guess building on the previous words

15. A friend says, 'AI knows exactly what it's going to say before it starts writing.' Why is this misleading?

    1. AI is too dumb to plan ahead
    2. AI decides each word one at a time based on probability, not from a pre-planned message
    3. AI always writes the same thing
    4. AI actually knows nothing