quiz bu lëkkale ak tegtal · Diggante Niveau

LLM-ni-àttekat Quiz

Xoolaat evaluation bu sukkandiko ci model, rubriques pointwise ak pairwise, àtte parçage ak validation ci tànneefi doomi aadama yi. ci jëf.

Yooni njiit yu jëm ciLLM ni àttekat
Laaj 1 de 8

Luy 'LLM-ni-àttekat'?

Xoolaatal laaj bu nekk8 laaj

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

  1. Luy 'LLM-ni-àttekat'?
    • Jëfandikoo benn xeetu làkk ngir joxe poñ wala méngale njariñu beneen xeetu làkk
    • Jëfandikoo misaalu làkk ngir bind àtte yoon
    • Royuwaay buy tontu waaw/déedet ci laaj yi
    • Modèle buy rang njariñu seetlu ci bis
  2. Lan mooy wuutale àtteb poñ ak àtteb ñaar?
    • Pointwise moo gëna gaaw; pairwise moo gëna yeex waaye nuru luko moy
    • Dañuy jox poñ benn tontu ci rubrik bi; ñaar ñaar tànn ki gëna baax ci ñaar
    • Pointwise du dox ludul ci kode; ñaar-ñaar rek ngir prose
    • Amul benn wuute dëgg
  3. Ci yii, yan ci ñoom mooy gëna ñaawe ci àttekati LLM yi?
    • Dama taamu tontu yuñ bind ak arafu mag
    • Saa yu nekk damay taamu baat bi njëkk ci alfabe
    • Taamu tontu yu gëna gudd, kalite bi du ci dara
    • Baña joxe poñ ci laaj yi jëm ci taarix
  4. naka la jàngat yi di faral di xeex position biais ci méngale ñaari mbir?
    • Sooy wane benn tontu rek saa yu nekk
    • Sooy njëkka def tontu bu gëna gudd
    • Soo efaasee barabu gëna jubal
    • Soo defee benn méngale ñaari yoon ak komànd biñ weccoo
  5. Ban anam jàngat la G-Eval jëfandikoo ngir defar mbind mi?
    • 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. Nan la ekip mëna firndeel ni àttekat LLM mingi méngoo ak tànneefi doomi aadama yi?
    • Buñu ko musa xoolaat ginaaw biñu ko tabb
    • Ni ñu ko kalibree ci set wurus buñ etikete nit
    • Defal àttekat bi note ci boppam
    • Soo ko jëfandikoo ci njeexte ayu-bis yi kese
  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