Practice from the curated 75 most-likely questions, or the full pool. Single- and multiple-answer items are graded exactly as written; drag-and-drop items show the reference answer to self-check against.
Where the two sources overlap, the curated set's answer is treated as correct.