Laaj ak tontu ci prefix

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

1. Lan mooy prefix tuning?
   • diisaayu transformatëru model bi yépp
   • du lenn ludul couche vocabulaire bi model bi di génne
   • Benn kopi bu wuute ci xeetu liggéey bu nekk
   • ab ensemble bu ndaw bu ay vecteur prefix yu wéy ci diir bi model bi di des
2. Fan lañu dugal vecteur prefix yi?
   • Ci biir caabi yi ak valeur yi ci bepp lalu bàyyi xel
   • Ci couche de sortie bi mujj kese
   • Ci biir fonction perte bi kese
   • Ci biir tokenizer bi balaa ngay dugal
3. Kan moo dugal prefix tuning ak kañ la?
   • 2018
   • Li ak Liang ci Stanford ci 2021
   • Google in 2015
   • Meta in 2023
4. Lan moo waral prefix tuning neex ngir liggéey lu bari?
   • Du soxla benn done tàggat yaram
   • Dafay tax inference dox ci benn hardware
   • Liggéey bu nekk soxlawul ludul ab prefix bu ndaw ci palaasu ab model bu mat sëkk
   • Dafay binndaat ba fàww xeetu liggéey yépp ci benn yoon
5. Ban pexe lañuy faral di jëfandikoo ngir dakkal tàggat prefix bi?
   • Yokk ñaari yoon limu jàng ci jéego bu nekk
   • Gelé prefix bi ak tàggat model bi
   • Dindi bépp layer buy bàyyi xel
   • Reparametriseer prefiks bi jaaraleko ci ab reso feed-forward bu ndaw buñuy sànni ci ganaw
6. Ban xeetu liggéey la prefix tuning gëna am solo?
   • Defar làkk wiñ nàmp lu melni tablo-ci-tekst ak tënk
   • Xeetu nataal yi
   • Synthese forme d'onde audio
   • Indexation de base de done
7. What proportion of model parameters did the original prefix-tuning paper report learning in its experiments?
   • About 10%
   • About 0.1%
   • Lu tollu ci 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