Skip to content

English · Español

Phase 18 — Quizzes (mirror)

🇪🇸 Las preguntas canónicas viven en data/quizzes/phase-18-training-loop.yaml. Este archivo es el espejo en Markdown para repaso rápido.

The source of truth is data/quizzes/phase-18-training-loop.yaml. The portal seeds quizzes from there. This page mirrors them for quick reading without spinning up the portal.


q-18-01 — AdamW decoupling: where does λθ enter?

Prompt (EN): In AdamW, where does the weight-decay term λ·θ_{t-1} enter the per-step update?

  • A. Added to the gradient g_t before the moment update.
  • B. Added to the parameter update, alongside the bias-corrected m̂ / (√v̂ + ε) term.
  • C. Multiplied into the learning-rate schedule.
  • D. Subtracted from v_t to control variance.

Correct: B. The "W" in AdamW is decoupled weight decay — the term enters the update, not the gradient. Adding it to g_t is Adam-L2, a different algorithm.


q-18-02 — Warmup duration

Prompt (EN): With lr_max = 3e-4, warmup_steps = 100, what is the learning rate at step 25 (linear warmup)?

  • A. 3e-4
  • B. 1.5e-4
  • C. 7.5e-5
  • D. 0

Correct: C. Linear warmup at step t gives lr_max · t / W = 3e-4 · 25 / 100 = 7.5e-5.


q-18-03 — Gradient clipping policy

Prompt (EN): Why use global L2-norm clipping instead of per-tensor norm clipping?

Free response. Expected mentions: preserves direction of the update; per-tensor changes the direction across parameters; global clip rescales uniformly.


q-18-04 — Which decay configurations are valid for an MLP block at §A13 scale?

Prompt (EN): Select every configuration that is reasonable for the AdamW param_groups of an MLP block at §A13 scale.

  • A. Decay applied to Linear weights with λ = 0.1.
  • B. Decay applied to LayerNorm scale parameters with λ = 0.1.
  • C. Decay applied to bias vectors with λ = 0.1.
  • D. No decay on biases or LN scale; λ = 0.1 on Linear / Embedding weights.

Correct: A, D. Decay on bias / LN scale (B, C) collapses these toward zero, hurting expressiveness without a reason — they aren't where overfitting lives.


q-18-05 — Bias correction omission

Prompt (EN): A learner reports that "warmup is too aggressive — even at step 50 the updates look tiny". They are using lr_max = 3e-4, warmup = 100. What is the most likely bug?

  • A. The warmup schedule is wrong.
  • B. The optimizer is using m_t and v_t directly without bias correction m̂_t = m_t/(1−β₁^t).
  • C. The clip threshold is too low.
  • D. The dataloader is shuffled per-step and never reaches the same example twice.

Correct: B. At small t, m_t ≈ (1 − β₁) · g₁ ≈ 0.1 g₁. Without bias correction, the first ~50 steps update at ~10% of intended magnitude — looks like aggressive warmup, is actually missing bias correction.