This lab reproduces one of the best-established findings in cognitive psychology: when people freely recall a studied list, accuracy traces a U-shaped curve across serial position — high at the start (primacy), high at the end (recency), and lowest in the middle. A row of word "pedestals" is presented in sequence at a rate you control; a rising bar over each pedestal then shows the model's predicted recall probability, driven by a dual-store account in which early items accumulate more rehearsal into long-term memory while late items are still resident in a fragile short-term buffer. Add a retention delay before recall and watch the recency bump collapse while primacy survives — the same dissociation Glanzer and Cunitz demonstrated with a real distractor task in 1966.