Text Representation Checkpoint

Reconstruct a new trace and diagnose implementation mistakes. Select one answer per multiple-choice question.


Get every question correct to pass this checkpoint. You can review feedback, revise your answers, and submit again.

1. Under this exercise tokenizer, what are the tokens of "Blue, BLUE"?

Select one answer.

2. Training documents are ["a a b", "b"]. What is document frequency DF(a)? Enter an integer.
3. A word appears in every document. Under the unsmoothed exercise IDF, what weight does it get?

Select one answer.

4. Vocabulary contains only red. What TF vector does compute_tf("red gold",vocab) return?

Select one answer.

5. You reorder vocabulary indices but keep the old IDF vector. What breaks?

Select one answer.

6. Two documents are "dog bites person" and "person bites dog". Their BoW rows match. What does that show?

Select one answer.

Join Discord