Kediska Qaabaynta Luuqadda Madow

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Waa maxay shaqada aasaasiga ah ee tababarka ee qaabaynta luqadda daboolan?
   • Odoroska erayada si badheedhka ah ee ka qarsoon macnaha guud
   • Odoroska kelmadda soo socota jumlad ka dib
   • Turjumaada qoraalka u dhexeeya laba luqadood
   • Weedho ku kala saraynta saxnaanta naxwaha
2. Qiyaastii boqolkiba meeqa calaamadood ayay BERT caadi ahaan maaskarada u qarisaa xilliga tababarka ka horreeya?
   • Ilaa 50%
   • Qiyaastii 15%
   • Ilaa 1%
   • Ilaa 90%
3. Muxuu MLM u bixiyaa qaabka fahamka laba jiho?
   • Waxay u akhridaa weedho gadaal
   • Waxay isticmaashaa laba nooc oo kala duwan
   • Calaamad kastaa waxay u dhigantaa ereyada bidix iyo midig labadaba
   • Kaliya waxay eegaysaa kelmad hore
4. Qorshaha waji-xidhka ee BERT, maxaa ku dhacaya ilaa 10% boosaska la doortay intii ay noqon lahaayeen [MASK]?
   • Waa la tirtiray gebi ahaanba
   • Waa la soo koobay
   • Waa la weyneeyay
   • Waxa lagu beddelaa kelmad aan toos ahayn
5. Meelkee lagu xisaabiyaa khasaaraha MLM?
   • Kaliya meelaha la daboolay
   • Calaamad kasta oo jumlada ah
   • Kaliya calaamadaha ugu horreeya iyo kan dambe
   • Kaliya calaamadaha xarakaynta
6. Marka la barbar dhigo saadaasha xigta-token, maxay MLM ka ganacsanaysaa?
   • Awoodda isticmaalka macnaha guud
   • Kartida soo saarida qoraal faseexa ah bidix-ilaa-midig
   • Helitaanka ereyada
   • Isticmaalka dareenka
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