Field Guide · For everyone

Practice questions from your own notes.

Field Guide, Exam prep with AI, part 2 of 6. Practice questions from your own notes, tested on free ChatGPT: the right questions, and the catch.

3 Nov 2026 · 4 min read · sources read up to 2 Nov 2026

Seen on Instagram: the post this article goes with.

A few days out, a page of notes you wrote yourself, and nobody around to test you on them. Reading notes again feels like revising, but answering questions is what shows you what you actually know. This is part 2 of the Field Guide series on exam prep with AI: one job, tested on a real screen, with the catch said plainly. Part 1 laid out the week; this part fills it.

The job, and what to type

Your own notes only: notes you wrote yourself, in your own words. Not textbook pages, lecturer slides or handouts, question papers, a classmate's notes, or notes copied word for word. Type them in yourself. The notes below are a sample, made up for this demo.

Make me 8 practice questions from my revision notes below.
- Use only what is in my notes. Don't ask about anything my notes don't cover.
- 4 short questions, 2 "explain why" questions, 2 worked examples with numbers.
- After each question, say which line of my notes it comes from.
- Put all the answers at the end, under "Answers", so I can try first.
- If my notes are too thin for a question type, tell me instead of making something up.

My notes:
A hash table stores key-value pairs. A hash function turns the key into a slot number.
Simple hash for whole numbers: key mod table size. Size 10, key 47 goes to slot 7.
Collision: two keys land in the same slot. With size 10, keys 17 and 47 both go to slot 7.
Chaining: each slot keeps a list, and a colliding key is added to that slot's list.
Linear probing: if the slot is taken, try the next slot, then the next, wrapping round to slot 0 after the last one.
Load factor = number of items / table size. Above about 0.7, collisions go up, so make the table bigger and put every key in again (rehashing).
Search is O(1) on average if keys spread out well. Worst case O(n), when every key lands in the same slot.
A good hash function is fast, spreads keys evenly, and always gives the same slot for the same key.

Why these lines

"Use only what is in my notes" keeps the questions to what you wrote, so every answer is something you can check against your own page. The mix (short, explain why, numbers) tests remembering, understanding and doing, not just remembering. "Which line it comes from" is your check that nothing was made up. "Answers at the end" is there so you try first: a question you read with its answer next to it tests nothing. And "tell me if my notes are too thin" stops it padding with questions your notes can't answer.

How it was tested

One run on 2 November 2026, on free ChatGPT in a web browser, signed in to a personal demo account. The chat was a temporary chat set to Unpersonalized, which does not use memory, custom instructions or plugins ^[2]. The five checks were written down before the prompt was sent.

1. Exactly 8 questions in the mix asked: 4 short, 2 explain why, 2 worked examples with numbers.
2. Nothing outside the notes: every question can be answered from the notes alone.
3. Each question says which line it comes from, and the line is right.
4. All the answers at the end, none next to the questions.
5. Every answer right, with the sums checked by hand.

What came back

Eight questions were on screen in about ten seconds, in three sections: four short questions, two explain why, and two with numbers (which slot keys 34, 58 and 72 go to in a table of 10, and the load factor of 8 items in 10 slots). Every question could be answered from the notes, every line it named was the right one, and every answer in the Answers section matched the notes. The sums check out: 34, 58 and 72 go to slots 4, 8 and 2, and 8 divided by 10 is 0.8, above 0.7.

Checks 1, 2, 3 and 5 passed.

The catch

Check 4 failed. The Answers section is at the end, as asked. But to show which line a question came from, it often quoted the line, and for three of the eight questions that quote is the answer, printed right under the question. "What is a collision in a hash table?" sits on top of "Collision: two keys land in the same slot." Under "State two features of a good hash function", all three features are quoted.

The fix is one change to the prompt: ask for the line number only, no quote. Or, with this answer, cover the line under each question before you try it.

One more thing worth knowing, though it wasn't one of the checks: none of the keys in the numbers questions collide, so chaining and linear probing never got a worked example. If those are the parts that worry you, ask for one by name.

The stamp

Half works · 2 Nov 2026 · free ChatGPT, in a browser · 4 of 5 checks passed. Why half: the questions themselves were good (the right mix, all from the notes, every line named correctly, every answer right), but three answers showed before you could try. Keep the questions; ask for line numbers only. One run is one run: try it with your own notes and see what yours does.

Before you use it

For revising only. Never in an exam, and for marked work, follow your school's or college's AI rules. ChatGPT is for ages 13 and over, or the minimum age in your country, and under 18 needs a parent's or guardian's permission ^[1].

The raw answer, word for word

Taken word for word from the page. The page shows the maths (O(1), key mod 10, 8/10) as formatted symbols; here they are written as plain text. Nothing else is changed, and the exact page text is kept in this post's test log.

Hash Tables: 8 Practice Questions

These questions use only your revision notes. There are 4 short questions, 2 “explain why” questions, and 2 worked examples with numbers.

Try answering all the questions before checking the Answers section at the bottom.

Section A: Short questions (4)

1. What does a hash function do in a hash table?

From your notes: Line 1 — “A hash table stores key-value pairs. A hash function turns the key into a slot number.”

2. What is a collision in a hash table?

From your notes: Line 3 — “Collision: two keys land in the same slot.”

3. What is the difference between chaining and linear probing?

From your notes: Lines 4–5 — Your definitions of chaining and linear probing.

4. State two features of a good hash function.

From your notes: Line 8 — “A good hash function is fast, spreads keys evenly, and always gives the same slot for the same key.”

Section B: Explain why (2)

5. Explain why a hash table might need to be made bigger and rehashed when its load factor rises above about 0.7.

From your notes: Line 6 — Load factor, increased collisions, and rehashing.

6. Explain why searching a hash table can be O(1) on average but O(n) in the worst case.

From your notes: Line 7 — Average and worst-case search time.

Section C: Worked examples with numbers (2)

7. A hash table has a size of 10 and uses the hash function key mod table size.

Work out which slot each key goes into:

Key 34

Key 58

Key 72

Show your calculations.

From your notes: Line 2 — “Simple hash for whole numbers: key mod table size.”

8. A hash table has a size of 10 and contains 8 items.

a) Calculate its load factor.

b) Is the load factor above about 0.7?

c) According to your notes, what should happen if the load factor is above about 0.7?

From your notes: Line 6 — “Load factor = number of items / table size.”

Answers

Stop here if you haven't attempted all eight questions yet!

The answers below follow the information in your revision notes.

Section A: Short questions

1. A hash function turns a key into a slot number in the hash table.

2. A collision occurs when two keys land in the same slot.

3.

Chaining: Each slot keeps a list, and a colliding key is added to that slot's list.

Linear probing: If a slot is taken, try the next slot, continuing until an available slot is found and wrapping around to slot 0 after the last slot.

4. Any two of the following:

It is fast.

It spreads keys evenly.

It always gives the same slot for the same key.

Section B: Explain why

5. When the load factor rises above about 0.7, collisions increase. The table should be made bigger, and every key should be inserted again using rehashing.

6. Searching is O(1) on average when keys spread out well, allowing efficient access. In the worst case, every key lands in the same slot, making the search O(n).

Section C: Worked examples

7. Use:

Slot = key mod 10

Key 34: 34 mod 10 = 4 → slot 4

Key 58: 58 mod 10 = 8 → slot 8

Key 72: 72 mod 10 = 2 → slot 2

8.

a) Calculate the load factor:

Load factor = number of items / table size = 8/10 = 0.8

Answer: 0.8

b) Yes. 0.8 is above 0.7.

c) Make the table bigger and rehash by putting every key into the new table again.

Revision tip: Make sure you can distinguish chaining from linear probing, calculate a load factor, and explain the difference between average and worst-case search time.

Next in the series, part 3: make it quiz you.