Item analysis

Question statistics

Attempts · 6Questions with data · 5

Question	n	Difficulty p	Discrimination r_pb	Verdict
What does SGD with a constant learning rate achieve on a convex loss function? r_pb=-0.12 < 0.15 — does not separate strong from weak	6	0.67	-0.12	low discrimination
Which of the following is true about the ReLU activation function? r_pb=-0.35 < 0.15 — does not separate strong from weak	6	0.83	-0.35	low discrimination
Why is forward‑mode automatic differentiation impractical for training large neural networks? r_pb=-0.12 < 0.15 — does not separate strong from weak	6	0.33	-0.12	low discrimination
What is the primary advantage of Adam over standard SGD with momentum?	6	0.33	0.32	OK
Why does standard momentum help in optimizing ill‑conditioned quadratics like x² + ρ y² (ρ ≪ 1)?	6	0.83	0.76	OK

Rejection rule (from 5 attempts): a question is flagged when its difficulty p falls outside 0.30–0.85 (share answering correctly) or its point-biserial discrimination drops below 0.15 (correlation with the rest of the attempt score). Retiring and regenerating flagged questions runs offline: uv run python -m app.quality … --regenerate.