Horgalayaasha Tuning Quiz

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Maxay dhab ahaantii tabobaraysaa hagaajinta horgalayaasha?
   • Dhammaan miisaannada beddelka moodeelka
   • Kaliya lakabka erayada wax soo saarka ee moodeelka
   • Nuqul gaar ah oo qaabka hawl kasta
   • Qayb yar oo horgaleyaal joogto ah oo horgale ah inta uu qaabku sii jiro baraf
2. Halkee lagu duray horgalayaasha horgalayaasha?
   • Geli furayaasha iyo qiyamka lakab kasta oo feejignaan ah
   • Kaliya lakabka ugu dambeeya ee wax soo saarka
   • Geli shaqada khasaaraha kaliya
   • Geli calaamadeeyaha ka hor inta aan la gelin
3. Yaa soo bandhigay hagaajinta horgalayaasha iyo qiyaas ahaan goorma?
   • OpenAI sanadka 2018
   • Li iyo Liang ee Stanford sanadka 2021
   • Google sanadka 2015
   • Meta sanadka 2023
4. Waa maxay sababta hagaajinta horgalayaasha u soo jiidasho leh u adeegida hawlo badan?
   • Uma baahna xog tababar haba yaraatee
   • Waxay ka dhigaysaa fikradda ku socota qalab la'aan
   • Hawl kastaa waxay u baahan tahay horgale yar oo keliya halkii ay ka ahaan lahayd nuqul buuxa
   • Waxay si joogto ah dib ugu qortaa moodalka saldhigga ee dhammaan hawlaha hal mar
5. Waa maxay khiyaamo inta badan loo isticmaalo in lagu dejiyo tababbarka horgalayaasha?
   • Laba-laabista heerka waxbarashada tallaabo kasta
   • Qaboojinta horgalayaasha iyo tababarida qaabka
   • Ka saarida dhammaan lakabyada dareenka
   • Dib-u-habaynta horgalayaasha iyada oo loo marayo shabakad yar oo gudbin ah oo hadhow la tuuray
6. Hawl noocee ah ayuu hagaajinta horgalayaasha si gaar ah u iftiimay?
   • Jiilka luqadda dabiiciga ah sida miis-ka-qoraal iyo soo koobid
   • Kala soocidda sawirka
   • Mawjada qaab maqal maqal ah
   • Tilmaanta database-ka
7. What proportion of model parameters did the original prefix-tuning paper report learning in its experiments?
   • About 10%
   • About 0.1%
   • Ilaa 50%
   • All model parameters
8. How did prefix-tuning perform in the paper’s low-data settings?
   • It always matched full fine-tuning on every task
   • It could not train with limited examples
   • It outperformed full fine-tuning in some evaluated low-data cases
   • It changed every backbone parameter