# Quiz with answers — Synthesis: a hybrid long-context architecture ### Question 1 Solve the worked-case variant: Remove MLA from the hybrid and replace it with a full KV cache. Trace the first effect on memory, retrieval fidelity, and latency without assuming a universal winner. - A. Per-token cache grows because keys and values are no longer compressed. Retrieval may retain more detail, but memory traffic and capacity rise. The verdict depends on quality, latency, and memory measurements on the actual workload. - B. Bounded design: periodic exact-attention layers for faithful retrieval, delta layers between them for fixed state, MLA to reduce per-token cache, MoE for conditional capacity, and spaced depth checkpoints. Verdict depends on quality/latency/memory measurements on the real task. - C. There is no universally best assembly. A diagram is a hypothesis; only controlled tests, hardware profiles, and user evaluations establish value. - D. That medium is the residual stream: text becomes tokens then embeddings, and one d_model = 4096 vector per position crosses the 48 blocks, each ADDING its contribution (session 20). The stream is the whole architecture’s data bus. **Answer: A.** Per-token cache grows because keys and values are no longer compressed. Retrieval may retain more detail, but memory traffic and capacity rise. The verdict depends on quality, latency, and memory measurements on the actual workload. 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 “Synthesis: a hybrid long-context architecture”? - A. Sequence mixing → Input and representation → Before inference - B. Input and representation → Before inference → Sequence mixing - C. Before inference → Input and representation → Sequence mixing - D. Before inference → Sequence mixing → Input and representation **Answer: C.** Before inference → Input and representation → Sequence mixing The chain follows the taught progression; reversing stages consumes a representation or state before it is produced. --- ### Question 3 If “Sequence mixing” is removed, which diagnostic method is defensible? - A. Also change the data to amplify the difference. - B. Keep the same input, predict the first output that depends on “Sequence mixing,” then compare the before/after trace. - C. Observe only the final output and invent the cause. - D. Conclude that the whole system fails before measuring. **Answer: B.** Keep the same input, predict the first output that depends on “Sequence mixing,” 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. One successful example proves the whole architecture is superior. - C. The mechanism name alone is enough for a production choice. - D. There is no universally best assembly. A diagram is a hypothesis; only controlled tests, hardware profiles, and user evaluations establish value. **Answer: D.** There is no universally best assembly. A diagram is a hypothesis; only controlled tests, hardware profiles, and user evaluations establish value. 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 “Synthesis: a hybrid long-context architecture”? - A. Mixed: established mechanisms + source-reported Kimi K3-style choices. - B. The route loads without an error. - C. Every learner opened the file. - D. The same result is assumed on every hardware target. **Answer: A.** Mixed: established mechanisms + source-reported Kimi K3-style choices. 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 “Depth and output”? - A. Never: the newest mechanism wins by default. - B. As soon as one memory metric falls, regardless of quality. - C. When a controlled test shows equivalent quality with lower memory, latency, or complexity. - D. As soon as the diagram contains fewer components. **Answer: C.** 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 “Input and representation.” Which remediation is most useful? - A. Accept the answer because the final number is correct. - B. Rebuild the first missing transformation, label its inputs and outputs, then test a neighboring case. - C. Provide the final result a second time. - D. Change several variables and ask for an intuition. **Answer: B.** 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 “Produce a bounded design verdict.”? - A. A list of terms without causal relations. - B. A screenshot without values or interpretation. - C. A confident claim without a baseline or threshold. - D. A trace with starting data, transformations, observed result, boundary, and next experiment. **Answer: D.** 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.