quiz bu lëkkale ak tegtal · Diggante Niveau

Xeetu neexal yi ci Process

Test bi dafay neexal xeetu poñ yi ci wàllu jéego, etiketu nit ak otomatik, ranking pexe, seetlu garab, ak lokalisasioŋ bu ñàkk njariñ.

Yooni njiit yu jëm ciModelu neexal ci liggéey
Laaj 1 de 8

Lan mooy njëkka wuutale xeetu neexal ci liggéey ak xeetu neexal ci njariñ?

Xoolaatal laaj bu nekk8 laaj

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

  1. Lan mooy njëkka wuutale xeetu neexal ci liggéey ak xeetu neexal ci njariñ?
    • Dafay joxe poñ ci jéego bu nekk ci diggante xalaat, du ci tontu bu mujj bi kese
    • Dafay gëna gaaw ci waxtu inference
    • Soxlawolen ay done tàggat
    • Du dox ludul ci dugal nataal
  2. Ban setu done bu siiw ci etiketu math yu nit ñi boole ci gëstu PRM?
    • Nataal Net
    • 800K PRM
    • SQUAD
    • Wikitekst
  3. Lan moo waral siñaalu neexal biy wane njariñ rek mën na réerloo nit?
    • Tontu yu mujj duñu musa jub
    • Royuwaayi njariñ mënu lire mbind
    • Tontu bu mujj bu jaar yoon mën na juddoo ci xalaat bu jaarul yoon, wala tontu bu juum ci xalaat bu gëna jub
    • Dafay jaar yoon saa yu nekk
  4. Benn anam buñu mëna jaar ngir soppi poñ yi ci jéego bu nekk ñu nekk benn poñ mooy:
    • Yokk benn nimero ci jéego bu nekk
    • Baña bàyyi xel ci jéego bu njëkk bi
    • Jëfandikool lim token yi kese
    • Jël jéego bi gëna ndaw ci wàllu probabilite ci chaine bi
  5. Lan la Math-Shepherd di jëfandikoo ngir etikete jéego yi ci saasi?
    • Monte Carlo dafay xayma ñaata yoon la benn jéego di yóbbe tontu bu jaar yoon
    • Xoolaat bu eksper bi def ci jéego bu nekk
    • Tekki jéego yi ci beneen làkk
    • Lim màrku ponksioŋ yi
  6. Ban risk moo gëna am solo sooy tàggat ay xeetu PRM?
    • Model bi fàtte na ni ñuy jàngee
    • Pirateri neexal, fu jéego yi di nuru lu baax ci PRM te duñu jub dëgg
    • PRM du lekk benn ordinatër
    • Jéego yi nekk nañu luñu mënatul defar
  7. How can a process reward model guide a tree search over reasoning?
    • Expand every branch equally without scoring
    • Choose branches by their final token length
    • Prioritize expanding partial solutions whose intermediate steps receive stronger scores
    • Replace all branches with one randomly selected answer
  8. What does the PRM800K dataset contain?
    • Only final answer labels for image captions
    • Audio recordings of spoken explanations
    • A corpus of code commits with no reasoning labels
    • Human labels for intermediate steps in mathematical reasoning solutions