Exercises

LM-C16-C-001

How many observed target tokens does one next-token prediction position have?

Compute it first, then check your number.

LM-C16-C-002

Does scale remove the need for evaluation? Answer 1 for yes or 0 for no.

Compute it first, then check your number.

LM-C16-C-003

If a document appears 3 times in a dataset, how many copies are present?

Compute it first, then check your number.

LM-C16-C-004

Should one benchmark be treated as final truth? Answer 1 for yes or 0 for no.

Compute it first, then check your number.

LM-C16-C-005

Can likely text be false? Answer 1 for yes or 0 for no.

Compute it first, then check your number.