Kuiz berkaitan panduan · Keras Tahap

Kuiz Pengesahan Persampelan Spekulatif

Menguji kebarangkalian pensampelan spekulatif, penerimaan dan pembetulan yang tepat, kos draf dan kekangan pada tuntutan prestasinya.

Laluan panduan yang berkaitanPengesahan Persampelan Spekulatif
soalan 1 daripada 8

Apakah idea teras di sebalik pensampelan spekulatif?

Pratonton setiap soalan8 soalan

Jawapan dan penjelasan muncul semasa anda mengambil kuiz.

  1. Apakah idea teras di sebalik pensampelan spekulatif?
    • Gunakan model draf kecil untuk mencadangkan berbilang token yang disahkan oleh model besar secara selari
    • Latih satu model gergasi untuk melangkau lapisan secara dinamik
    • Cache setiap ayat yang mungkin lebih awal
    • Gantikan pensampelan dengan penyahkodan tamak untuk menjimatkan masa
  2. Mengapa pensampelan spekulatif tidak merendahkan kualiti keluaran?
    • Kerana model draf sentiasa betul
    • Kerana peraturan pengesahan menjadikan pengedaran akhir sama dengan pensampelan sasaran tulen
    • Kerana ia hanya berfungsi pada gesaan pendek
    • Kerana ia purata output kedua-dua model
  3. In exact speculative sampling, what happens when a drafted token is rejected?
    • It outputs the rejected draft token unchanged
    • It discards the entire conversation and restarts
    • The algorithm samples from a corrected residual distribution at that position
    • It permanently changes the target model
  4. Under exact speculative sampling, what is guaranteed about the output distribution?
    • It is always deterministic
    • It matches the draft model’s distribution instead
    • It is more diverse than every target-model output
    • It matches the target model’s sampling distribution
  5. What information does the target verification pass provide in exact speculative sampling?
    • Target probabilities for proposed positions, used for acceptance tests and residual resampling
    • A guarantee that no target-model computation is needed
    • Only the draft model’s next-token distribution
    • A corrected token sequence without any sampling
  6. Which description correctly distinguishes two self-speculative approaches?
    • Both use only an unrelated small draft model
    • Medusa adds prediction heads, while EAGLE uses feature-space proposals
    • Medusa is a tokenizer and EAGLE is a reward model
    • Both remove the target model from decoding
  7. The draft assigns probability 0.8 to a token and the target assigns 0.4. What is the speculative acceptance probability?
    • 0.2
    • 0.4
    • 0.5
    • 1.0
  8. What most limits speculative sampling’s realized speedup?
    • The color of the model’s interface
    • The target model’s name length
    • Whether the prompt contains punctuation
    • The draft-model cost and how much proposed work the target accepts