Kedis ku xidhan hagaha · Dhexdhexaad Heerka

Kediska Xulashada Xulashada Khatarsanaanta

Sharaxaayaa ORPO's tixraaca-qaabka doorbidida ikhtiyaarka bilaashka ah, ciqaabta saamiga-khakhuska, iyo tababarka la kormeerayo iyo doorbidida la isku daray.

Wadooyinka haga ee la xidhiidhaKordhinta Xulashada Saamiga Khasaarada
Su'aal 1 ee 8

Waa maxay faa'iidada ugu weyn ee ORPO ee hababka sida DPO?

Horudhac su'aal kasta8 su'aalo

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

  1. Waa maxay faa'iidada ugu weyn ee ORPO ee hababka sida DPO?
    • Waxay meesha ka saaraysaa baahida loo qabo nooc tixraaceed oo gaar ah
    • Uma baahna xog tababar haba yaraatee
    • Waxay kaliya ku shaqeysaa moodooyinka sawirka
    • Waxay dammaanad qaadaysaa muuqalka eber
  2. Waa maxay isbarbardhigga saamiga 'khasaaraha' ee ORPO?
    • Xawaaraha tababarka iyo saxnaanta
    • Fursadaha jawaabta la doortay iyo jawaabta la diiday
    • Laba heerar waxbarasho oo kala duwan
    • Cabbirka moodeelka iyo cabbirka xogta kaydinta
  3. What form does an ORPO preference example contain?
    • Only a single unlabeled response
    • A prompt with two identical completions
    • A prompt with one chosen response and one disfavored response
    • A reward score with no associated responses
  4. How does ORPO convert a response probability p into odds?
    • One minus p divided by p
    • The square root of p
    • The logarithm of the prompt length
    • p divided by one minus p
  5. Khasaaraha ORPO ayaa sida ugu wanaagsan loogu sifeeyay isku darka?
    • SFT cross-entropy oo lagu daray erey doorbidida saami qaybsiga miisaan culus
    • Kaliya abaal-marinta buundada
    • Soo laabasho waxbarasho oo xoojin ah oo saafi ah
    • Cirbad qaylada random
  6. Which response receives the supervised fine-tuning component of ORPO?
    • Only the rejected response
    • The chosen or preferred response
    • A response sampled after deployment
    • No response; ORPO has no language-modeling term
  7. Which parameter controls the weight of ORPO’s preference term relative to the supervised loss?
    • Cabbirka ereyada tokenizer
    • Tirada madaxda dareenka
    • The preference-loss coefficient
    • The prompt’s character count
  8. Why should an ORPO result report its preference data and evaluation setup?
    • ORPO results are identical across datasets
    • The objective removes the need for evaluation
    • The number of model layers determines every preference
    • Outcomes depend on the dataset, coefficient, base model, and evaluation