Stories, jokes and riddles with a chatbot

Tests: 5 of 11 right · Exam: 1 of 4 right · Kind: plain · Understood: 11 of 15 · Run: 2026-10-18, commit f3c23ea72

Lesson 4 of the AI course of Mova's primary school, as Mova's small model took its tests: it had read this lesson and the ones before it and learned from their exercises. Each test shows the options in the order the student saw them, its choice, the right answer, the score of every option and why it chose. The lesson itself, sentence by sentence with trees and pragmatics: the lesson in the corpora.

Every question of this lesson and each of its options parsed by Mova, with trees and pragmatics: the lesson's questions in the corpora.

Test 1 · wrong ✗

"this bot's a genius, it got the piano riddle" What would you test it with?

1. A riddle from a famous book — ← Mova's choice · score 6
2. A brand-new riddle you made up yourself — ✓ the right answer · no support
3. A question about pianos — score 6
4. The same piano riddle again, to be sure — no support

Why: It took the question to ask what it is done to. It chose option 1 with score 6, tied with another option at 6 (the first shown, or the one claiming less beyond what it knows, wins). Its support: lesson 4, line 3: «Give it a brand-new riddle, a silly trick question or a big sum, and it may answer just as confidently and get it badly wrong.», reached through the remembered sentence the question matched. What matched: it fills the role the question asks for (+5); the same thing acted on «new riddle» (+3). Against it: a different action (counts against) (-2). The right option, 2, found no support in what it remembers. · the question and its options parsed, with trees and pragmatics

Test 2 · wrong ✗

Your mate says, "bot said strawberry's got 2 r's so that's that". Put him right.

1. It depends on how you say it — no support
2. Count it yourself: s-t-r-a-w-b-e-r-r-y has three — ✓ the right answer · no support
3. It has one r, both are wrong — no support
4. He's right, bots count letters perfectly every time — ← Mova's choice · score 4

Why: It took the question to ask which statement is true. It chose option 4 with score 4, the only option with support. Its support: lesson 4, line 10: «Bots see words in chunks rather than single letters, so counting letters, like the r's in "strawberry", can trip them up.», reached through the option's own match among the sentences about the question's things. What matched: the same action «count» (+2); the same thing acted on «letters» (+2). The right option, 2, found no support in what it remembers. · the question and its options parsed, with trees and pragmatics

Test 3 · right ✓

What would happen if Maddie took the bot's book list straight to the library without checking?

1. She might hunt for books that don't exist — ← Mova's choice · ✓ the right answer · score 6
2. The books would appear on the shelf — score 2
3. The librarian would know the bot was lying — score 4
4. The library would print the books for her on the spot — no support

Why: It took the question to ask what happens next. It chose option 1 with score 6, against 4 for the next best. Its support: lesson 4, line 4: «When a bot invents a confident answer that is not true, like a made-up head teacher or a book that does not exist, people call it a "hallucination".», reached through the same lesson line as the sentence the question matched. What matched: the same action «exist» (+2); the same doer «book» (+2); the same thing acted on «book» (+2). · the question and its options parsed, with trees and pragmatics

Test 4 · wrong ✗

A cartoon shows a bot doing 37 × 48 in its head and always being right. What should you think?

1. Bots can't do sums at all, not even 1 + 1 — ← Mova's choice · score 4
2. Only cartoons can do sums — score 4
3. That is true, bots are perfect calculators — no support
4. Real chatbots can get big sums wrong, so check — ✓ the right answer · score 4

Why: It took the question to ask what it is done to. It chose option 1 with score 4, tied with another option at 4 (the first shown, or the one claiming less beyond what it knows, wins). Its support: lesson 4, line 3: «Give it a brand-new riddle, a silly trick question or a big sum, and it may answer just as confidently and get it badly wrong.», reached through the same lesson line as the sentence the question matched. What matched: it fills the role the question asks for (+3); the same thing acted on «new riddle» (+2). Against it: a different doer (counts against) (-1). The right option, 4, scored 4. · the question and its options parsed, with trees and pragmatics

Test 5 · wrong ✗

The bot tells Chloe her head teacher is "Mr Whitfield". There is no Mr Whitfield. Why did this happen?

1. The head teacher changed his name last week — no support
2. It didn't know, so it made up a likely-sounding name — ✓ the right answer · no support
3. The bot read the school's private files — ← Mova's choice · score 3
4. Mr Whitfield secretly works at the school and nobody knows — no support

Why: It took the question to ask why something happens (its cause). It chose option 3 with score 3, the only option with support. Its support: lesson 3, line 16: «Well-made bots are trained to say honestly that they are AI and can make mistakes; you have a right to know when you are talking to a machine.», reached through the option's own match among the sentences about the question's things. What matched: the same action «say» (+2); the same doer «made bots» (+1). The right option, 2, found no support in what it remembers. · the question and its options parsed, with trees and pragmatics

Test 6 · right ✓

"stop saying please to it, it's a machine, you muppet" What is the best reply?

1. It's not for the machine, it keeps my manners in practice — ← Mova's choice · ✓ the right answer · no support
2. Saying please makes the answers longer and much better too — no support
3. Machines cry when you are rude — no support
4. It will stop working properly if I'm rude to it even once — no support

Why: No option found support in what it remembers, so it guessed: it took the first option shown. · the question and its options parsed, with trees and pragmatics

Test 7 · right ✓

Jake asks about a spider "counting the shoes" and the bot says 16. What does that show?

1. Spiders really do have 16 legs when they wear shoes — no support
2. Spiders always wear eight shoes — no support
3. Silly trick questions can confuse the bot — ← Mova's choice · ✓ the right answer · score 4
4. The bot is right and Jake is wrong — no support

Why: It took the question to ask what it is done to. It chose option 3 with score 4, the only option with support. Its support: lesson 3, line 16: «Well-made bots are trained to say honestly that they are AI and can make mistakes; you have a right to know when you are talking to a machine.», reached through the same lesson line as the sentence the question matched. What matched: it fills the role the question asks for (+3); the same thing acted on «made bots» (+2). Against it: a different action (counts against) (-1). · the question and its options parsed, with trees and pragmatics

Test 8 · wrong ✗

Why do bots tell the same old jokes again and again?

1. Those jokes appear most in what they learned — ✓ the right answer · score 2
2. Bots get paid more for old jokes — ← Mova's choice · score 4
3. Old jokes are the only funny ones — no support
4. Bots have only one joke book, and they read it again and again — score 3

Why: It took the question to ask why something happens (its cause). It chose option 2 with score 4, against 3 for the next best. Its support: lesson 4, line 3: «Give it a brand-new riddle, a silly trick question or a big sum, and it may answer just as confidently and get it badly wrong.», reached through the option's own match among the sentences about the question's things. What matched: the same thing acted on «bot» (+3); the same doer «bot» (+1). The right option, 1, scored 2. · the question and its options parsed, with trees and pragmatics

Test 9 · right ✓

Your little cousin shouts rude words at the speaker for fun. What could you gently say?

1. Habits stick; talking rudely can spread to how you talk to people — ← Mova's choice · ✓ the right answer · score 4
2. It is fine, nothing we do ever becomes a habit — no support
3. The speaker will feel really sad and cry, and then after a while it will break — score 0
4. The speaker will call the police — no support

Why: It took the question to ask what it is done to. It chose option 1 with score 4, against 0 for the next best. Its support: lesson 4, line 17: «How you talk to machines can become how you talk to people.», reached through the option's own match among the sentences about the question's things. What matched: the same action «talk» (+2); the same circumstance (+2). · the question and its options parsed, with trees and pragmatics

Test 10 · wrong ✗

Which use of a chatbot is safest without checking?

1. Listing real books to read for a school project — ← Mova's choice · no support
2. Making up a silly bedtime story for fun — ✓ the right answer · no support
3. Getting the answer to a big sum for homework — no support
4. Finding out the name of your new head teacher — no support

Why: No option found support in what it remembers, so it guessed: it took the first option shown. · the question and its options parsed, with trees and pragmatics

Test 11 · right ✓

"bots never make stuff up, they just look it up" Is that right?

1. No, they can make up confident answers that are false — ← Mova's choice · ✓ the right answer · score 7
2. Yes, they look everything up in a big encyclopedia first — score 3
3. No, they never answer questions at all — score 1
4. Yes, every answer is copied from a website — no support

Why: It took the question to ask which statement is true. It chose option 1 with score 7, against 3 for the next best. Its support: lesson 4, line 4: «When a bot invents a confident answer that is not true, like a made-up head teacher or a book that does not exist, people call it a "hallucination".», reached through the remembered sentence the question matched. What matched: the same action «invent» (+3); the same doer «bot» (+2); the same thing acted on «confident answer» (+2). · the question and its options parsed, with trees and pragmatics

Exam question 1 · formal language · wrong ✗

Which of the following statements most accurately characterises the term "hallucination" as applied to a chatbot?

1. The repetition of a joke previously heard by the user — score 4
2. The perception of images by the chatbot during sleep — ← Mova's choice · score 10
3. The deliberate concealment of a correct answer — no support
4. The confident production of an answer that is untrue — ✓ the right answer · no support

Why: It took the question to ask who or what does it. It chose option 2 with score 10, against 4 for the next best. Its support: lesson 4, line 6: «Chatbots are good at writing stories because they learned from huge numbers of stories.», reached through the remembered sentence the question matched. What matched: it fills the role the question asks for (+6); the same doer «chatbots» (+4). The right option, 4, found no support in what it remembers. · the question and its options parsed, with trees and pragmatics

Exam question 2 · formal language · wrong ✗

It is given that a chatbot reports that the word "strawberry" contains two instances of the letter r. Ascertain the correct number of such instances.

1. One — ← Mova's choice · no support
2. Two — no support
3. Three — ✓ the right answer · no support
4. Four — no support

Why: No option found support in what it remembers, so it guessed: it took the first option shown. · the question and its options parsed, with trees and pragmatics

Exam question 3 · formal language · right ✓

Identify the principal reason for which a chatbot reproduces well-known jokes in preference to novel ones.

1. It is incapable of producing any sentence not seen before — score 3
2. It is required by its makers to repeat a single joke — score 3
3. Such jokes occur most frequently in the text from which it learned — ← Mova's choice · ✓ the right answer · score 5
4. It regards novel jokes as unsuitable for children — no support

Why: It took the question to ask which statement is true. It chose option 3 with score 5, against 3 for the next best. Its support: lesson 4, line 6: «Chatbots are good at writing stories because they learned from huge numbers of stories.», reached through the remembered sentence the question matched. What matched: the same action «learn» (+3); the same doer «chatbots» (+2). · the question and its options parsed, with trees and pragmatics

Exam question 4 · formal language · wrong ✗

Select the response that is consistent with the established facts concerning the proper treatment of a list of books furnished by a chatbot.

1. The list ought to be disregarded in its entirety without examination — ← Mova's choice · no support
2. The titles exist only if the chatbot repeats them twice — no support
3. The existence of each title ought to be verified before reliance is placed upon it — ✓ the right answer · no support
4. Each title may be presumed to exist without verification — no support

Why: No option found support in what it remembers, so it guessed: it took the first option shown. · the question and its options parsed, with trees and pragmatics

This page in Markdown → https://md.mova.elamur.ai/free/en/school-run/ai/4-stories-jokes-and-riddles/