# Quiz with answers — Chunking, causality, and parallel prefill

### Question 1

Solve the worked-case variant: Move from 8 to 10 tokens with chunk size 4. Draw the chunks, local causal masks, and states carried between chunks.

- A. For 8 tokens in chunks of 4, the first computes a 4×4 triangle then passes S₄. The second receives S₄, computes its local triangle, and produces S₈. No token in the first block can read the second.
- B. Chunking improves execution; it does not automatically change the information capacity of state.
- C. Chunk sizes are [4,4,2]. The first two use 4×4 triangles; the last uses a 2×2 triangle. End states S₄ then S₈ cross boundaries; no position may read a future token.
- D. The precise diagnosis: the DEPENDENCY is sequential (each S_t depends on S_{t−1}), but most of the per-token COMPUTE — local q·k products, k vᵀ writes — is not. Naive recurrence serializes everything because it never separates the two.

**Answer: C.** Chunk sizes are [4,4,2]. The first two use 4×4 triangles; the last uses a 2×2 triangle. End states S₄ then S₈ cross boundaries; no position may read a future token. The correct answer executes the requested change and gives a checkable result; the other texts do not close this calculation or trace.

---

### Question 2

Which causal order correctly connects the first three stages of “Chunking, causality, and parallel prefill”?

- A. Prefill and decode → Naive recurrence → Split into chunks
- B. Split into chunks → Naive recurrence → Prefill and decode
- C. Naive recurrence → Prefill and decode → Split into chunks
- D. Prefill and decode → Split into chunks → Naive recurrence

**Answer: A.** Prefill and decode → Naive recurrence → Split into chunks The chain follows the taught progression; reversing stages consumes a representation or state before it is produced.

---

### Question 3

If “Split into chunks” is removed, which diagnostic method is defensible?

- A. Also change the data to amplify the difference.
- B. Observe only the final output and invent the cause.
- C. Conclude that the whole system fails before measuring.
- D. Keep the same input, predict the first output that depends on “Split into chunks,” then compare the before/after trace.

**Answer: D.** Keep the same input, predict the first output that depends on “Split into chunks,” then compare the before/after trace. One intervention and a prior prediction make the delta attributable to the removed mechanism.

---

### Question 4

Which verdict respects this session’s validity boundary?

- A. The mechanism guarantees accuracy, speed, and stability for every workload.
- B. Chunking improves execution; it does not automatically change the information capacity of state.
- C. One successful example proves the whole architecture is superior.
- D. The mechanism name alone is enough for a production choice.

**Answer: B.** Chunking improves execution; it does not automatically change the information capacity of state. The correct answer bounds the conclusion; the others turn a local relation into a global guarantee.

---

### Question 5

Which evidence best matches the stated status of “Chunking, causality, and parallel prefill”?

- A. The route loads without an error.
- B. Every learner opened the file.
- C. Established mechanisms; numerical simplifications are pedagogical.
- D. The same result is assumed on every hardware target.

**Answer: C.** Established mechanisms; numerical simplifications are pedagogical. Product or mechanism evidence must remain attributed and measured; availability and completion do not prove value.

---

### Question 6

When should a simpler baseline be preferred to “Incoming and outgoing state”?

- A. When a controlled test shows equivalent quality with lower memory, latency, or complexity.
- B. Never: the newest mechanism wins by default.
- C. As soon as one memory metric falls, regardless of quality.
- D. As soon as the diagram contains fewer components.

**Answer: A.** When a controlled test shows equivalent quality with lower memory, latency, or complexity. The choice depends on a measured trade-off on the real workload, not novelty or one isolated metric.

---

### Question 7

A learner gets the right result but cannot explain “Naive recurrence.” Which remediation is most useful?

- A. Accept the answer because the final number is correct.
- B. Provide the final result a second time.
- C. Change several variables and ask for an intuition.
- D. Rebuild the first missing transformation, label its inputs and outputs, then test a neighboring case.

**Answer: D.** Rebuild the first missing transformation, label its inputs and outputs, then test a neighboring case. The remediation targets the first causal break and then requires transfer instead of rewarding a guessed result.

---

### Question 8

Which submission actually demonstrates the outcome “Explain causal triangular structure.”?

- A. A list of terms without causal relations.
- B. A trace with starting data, transformations, observed result, boundary, and next experiment.
- C. A screenshot without values or interpretation.
- D. A confident claim without a baseline or threshold.

**Answer: B.** A trace with starting data, transformations, observed result, boundary, and next experiment. The correct submission makes the reasoning reproducible and the verdict revisable by future measurement.