quiz.json

{
  "lesson": "06-sglang-radixattention",
  "title": "SGLang and RadixAttention for Prefix-Heavy Workloads",
  "questions": [
    {
      "stage": "pre",
      "question": "What core data structure backs SGLang's KV cache reuse?",
      "options": [
        "A radix tree where each node owns a token range and its KV blocks",
        "A hash table keyed by full prompt",
        "A skip list keyed by request id",
        "A B-tree of attention scores"
      ],
      "correct": 0,
      "explanation": ""
    },
    {
      "stage": "check",
      "question": "Why is FCFS scheduling wrong for prefix-heavy traffic on SGLang?",
      "options": [
        "FCFS only works on AMD GPUs",
        "FCFS is the recommended SGLang policy",
        "FCFS is incompatible with PagedAttention",
        "FCFS can evict a hot prefix branch before the next long-prefix request hits, breaking radix-tree reuse"
      ],
      "correct": 3,
      "explanation": ""
    },
    {
      "stage": "check",
      "question": "What eviction granularity does SGLang's cache-aware scheduler use to match radix shape?",
      "options": [
        "Per-request only",
        "Whole branches, starting from shortest-used leaves",
        "Single tokens",
        "Random eviction"
      ],
      "correct": 1,
      "explanation": ""
    },
    {
      "stage": "check",
      "question": "What is the most direct engineer's lever for keeping the radix-tree shared prefix discoverable?",
      "options": [
        "Lower the GPU memory utilization knob",
        "Fix prompt-template ordering so immutable content (system, tools, schemas) is always first",
        "Always use static batching",
        "Disable continuous batching"
      ],
      "correct": 1,
      "explanation": ""
    },
    {
      "stage": "post",
      "question": "Which workload pattern does the lesson NOT expect RadixAttention to win on?",
      "options": [
        "Agents with shared tool schemas",
        "Voice workloads with repeated preambles",
        "RAG with a shared retrieval preamble",
        "Single-shot generation with unique prompts and no shared system prompt"
      ],
      "correct": 3,
      "explanation": ""
    },
    {
      "stage": "post",
      "question": "ProjectDiscovery's deployment moved from 7% to 74% prefix-cache hit rate by doing what?",
      "options": [
        "Moving dynamic content out of the cacheable prefix",
        "Increasing GPU count from 8 to 16",
        "Switching from vLLM to SGLang without any prompt changes",
        "Disabling continuous batching"
      ],
      "correct": 0,
      "explanation": ""
    }
  ]
}