Ikibazo gihuza ibibazo · Hagati Urwego

LLM-nk-umucamanza Ikibazo

Ongera usuzume icyitegererezo gishingiye ku cyerekezo, icyerekezo kimwe na rubrics, guca urubanza kubogama no kwemeza ibyo abantu bakunda. mu myitozo.

Inzira zijyanye no kuyoboraLLM nk'umucamanza
Ikibazo 1 Bya 8

'LLM-nk-umucamanza' bivuga iki?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. 'LLM-nk-umucamanza' bivuga iki?
    • Gukoresha icyitegererezo cyururimi kugirango utange amanota cyangwa ugereranye indi moderi ibisubizo
    • Gukoresha icyitegererezo cyururimi kugirango wandike ibyemezo byemewe n'amategeko
    • Icyitegererezo gisubiza gusa yego / ntakibazo
    • Icyitegererezo gikurikirana ibisubizo byubushakashatsi kumatariki
  2. Ni irihe tandukaniro riri hagati yo guca imanza no guhuza ibitekerezo?
    • Icyerekezo cyihuta; byombi bitinda ariko birasa ukundi
    • Icyerekezo cyerekana icyiciro kimwe kirwanya rubric; byombi bitoranya ibyiza bibiri
    • Icyerekezo gikora kode gusa; byombi kubwa prose gusa
    • Nta tandukaniro nyaryo
  3. Ninde muribi urimo kubogama neza kubacamanza ba LLM?
    • Guhitamo ibisubizo byanditse mu nyuguti nkuru
    • Buri gihe uhitamo ijambo ryambere mubisobanuro
    • Gukunda ibisubizo birebire utitaye ku bwiza
    • Kwanga gutanga amanota kubibazo byamateka
  4. Nigute abasesengura barwanya imyanya ibogamye mugereranya?
    • Mugaragaza gusa igisubizo kimwe icyarimwe
    • Mugushira buri gihe igisubizo kirekire
    • Mugusiba ikibanza cyo gutsindishirizwa
    • Mugukoresha buri kugereranya kabiri hamwe na gahunda yahinduwe
  5. What evaluation procedure did G-Eval use for generated text?
    • Chain-of-thought steps followed by a form-filling evaluation against criteria
    • A speaker verification threshold
    • A search index with BM25 only
    • An image classifier trained on screenshots
  6. Nigute itsinda ryemeza ko umucamanza wa LLM akomeza guhuza nibyo abantu bakunda?
    • Ntabwo wongeye kugenzura nyuma yo gushiraho
    • Muguhinduranya kurutonde rwumuntu wanditseho zahabu
    • Mugukora urwego rwumucamanza ubwarwo
    • Mugukoresha gusa muri wikendi
  7. What is self-preference bias in LLM-based evaluation?
    • A judge always selects the shortest answer
    • A judge cannot compare two outputs
    • A judge favors outputs generated by its own model or model family
    • A judge evaluates only audio signals
  8. What does G-Eval use to make its discrete score more fine-grained in the reported method?
    • The position of the answer in a pairwise comparison
    • The judge model’s parameter count only
    • A count of citations in the answer
    • Probability weights over the candidate score tokens