Quiz modeling làkk masqué

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

1. Lan mooy liggéeyu tàggat yaram bu mag bi ci modeling làkk buñ maskeer?
   • Xalaatal kàddu yuñ nëbb ci anam wu leen wër
   • Xalaatal baat bi ci topp ginaaw ab tegtal
   • Tekki mbind diggante ñaari làkk
   • Raŋ santaane ci njubte ci nàfar
2. Lu tollu ci ban pursàntaasu jeton la BERT di faral di maskeer ci diiru tàggat yaram?
   • Lu tollu ci 50%
   • Lu tollu ci 15%
   • Lu tollu ci 1%
   • Lu tollu ci 90%
3. Lu tax MLM di joxe xeetu xam-xam bu am ñaari yoon?
   • Dafay jàng tegtal yi ci ginaaw
   • Dafay jëfandikoo ñaari xeetu mbir yu wuute
   • Token bu nekk dafay bàyyi xel ci kàddu yi ci cammoy ak ci ndeyjoor
   • Dafay xool baat bi njëkk kese
4. Ci xeetu maskeer bu BERT, lan mooy dall lu tollu ci 10% ci liggéey yiñ tànn ludul nekk [MASK]?
   • Dañu leen efaase lëmm
   • ñu ngi leen ñaari yoon
   • Ñu ngi leen bind ak baat yu mag
   • Dañu leen di wecci ak baat bu bari
5. Ci ban position lañuy xayma perte MLM bi?
   • Position yiñ maske rek
   • Bépp màndarga ci tegtal bi
   • Token yu njëkk yi ak yu mujj yi kese
   • Jeton ponksioŋ kese
6. Buñu ko méngale ak xayma token bi ci topp, lan la MLM di jaay?
   • Mën jëfandikoo contexte
   • Mën defar mbind mu neex cammoñ dem ndeyjoor
   • Yokk ci ab vokabileer
   • Jëfandikoo bàyyi xel
7. How does RoBERTa’s dynamic masking differ from BERT’s original setup?
   • It masks only the first token of every sentence
   • It predicts future tokens with a causal decoder
   • The masked positions can be regenerated across training passes instead of remaining fixed
   • It replaces all masked tokens with random numbers
8. What auxiliary objective did the original BERT pretraining recipe combine with MLM?
   • Image patch reconstruction
   • Audio speaker diarization
   • A reward model for code generation
   • Next-sentence prediction