quiz bu lëkkale ak tegtal · Diggante Niveau

Test-Time Xayma Eskalaasioŋ Quiz

Xoolaat xayma ci diiru inference, seetlu kàndidaa, xaaj bu méngoo ak njaay ak yeexal njëg yiñ natt ci njàngum diiru test.

Yooni njiit yu jëm ciJamonoy test
Laaj 1 de 8

Luy 'test-time calcul'?

Xoolaatal laaj bu nekk8 laaj

Sooy def quiz bi, tontu yi ak leeral yi dañuy feeñ.

  1. Luy 'test-time calcul'?
    • xayma jëfandikoo ngir tàggat model bi ci done
    • Jamono jiy jël ngir yebbi benn model
    • xayma biñu def su model bi tontu ab laaj ci inference
    • Kuuraŋ bi ñuy jëfandikoo ngir seddal santu done yi
  2. In test-time scaling studies, what can extra inference compute be used for?
    • Changing the model’s training data during the request
    • Increasing the number of model parameters automatically
    • Guaranteeing correctness without evaluation
    • Searching or revising candidate solutions and checking their quality
  3. Lan mooy xéewal gi gëna am solo ci diiru test-time escalier?
    • Jaar-jaar bu gëna mag ci weccoo latency bu gëna mag ak njëg
    • Memoire bu néew waaye tàggat yaram bu yeex
    • Done yu bari waaye model yu gëna ndaw
    • Grafik yu gëna baax ngir dayo bu woyof
  4. Luy 'gën-ci-N' échantillonnage?
    • Taggat ci N done yi gëna mag
    • Defar N tontu kàndidaa ak tànn bi gëna baax
    • Jëfandikool N baat yi njëkk ci benn laaj
    • Doxal model bi ci N ordinatër yu wuute ngir gaawaay
  5. Lan moo waral ñuy jàppee xayma waxtu test ni 'axe scaling' bu bees?
    • Dafay wecci bépp tàggat buñu soxla
    • ci model yu ndaw rek lay dox
    • Dafay yokk performance ci yokk inference compute moo gën yokk dayo model bi
    • Dafay wàññi limu jëfandikukat yi benn model mëna jàppale
  6. Ban resultaa bu yéeme la test-time bi mëna joxe?
    • Model yi mën nañu daw te duñu am benn GPU
    • Done tàggat yaram nekkul lu am njariñ
    • Model yi ñoo gëna gaaw ci tontu
    • Benn model bu gëna ndaw di xalaat lu gëna yàgg mën na xëcc model bu gëna mag di tontu ci saasi
  7. What is one reason a fixed best-of-N strategy may waste inference compute?
    • It spends the same candidate budget on easy and difficult prompts
    • It modifies model weights on every query
    • It cannot generate multiple candidates
    • It eliminates the need for scoring
  8. What did the test-time compute paper compare with best-of-N sampling?
    • A strategy that disables inference
    • A compute-optimal strategy that allocates effort based on expected solution quality
    • A larger tokenizer vocabulary
    • A method that changes image resolution