Quiz buy fàtte musiba

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

1. Ban anam mooy màndargaal fàtte bu metti ci reso neuronal?
   • Sudee jàng liggéey bu bees dafay tax model bi ñàkka mëna jàng bu njëkk
   • Su benn model jeexee ci mémoire bi ci diiru inference
   • Suñu efaasee done tàggat yaram ci anam wu jaarul yoon
   • Su model bi bañee tontu benn laaj
2. Lan moo waral fàtte bu metti di am ci reso neuronal yi?
   • Ndax liggéey bu nekk dafay am reso boppam ci saasi
   • Ndax yeesali gradient yu bees yi dañuy soppi poid yiñ bokk yuñ kode xam-xam yu yàgg yi
   • Ndax model yi mënu ñu denc lu ëpp benn mbir
   • Ndax ay couche attention dañuy efaase ay done ci luñu ko soobe
3. Lan lay def ngir wàññi fàtte?
   • Dafay efaase done yi ci liggéey bu bees bi
   • Yokk ñaari yoon limu jàng
   • Dafay yar coppite yi am ci poid yi am solo ci liggéey yi ñu jàngoon bu njëkk
   • Dindi bàyyi xel ci model bi
4. naka la ko jaar-jaar (repetition) di xeexe fàtte?
   • Sooy gelé model bi yépp ba fàww
   • Soo tàggatee ci done yu bees yi kese
   • Soo wàññi limu diisaay yi
   • Ni ñuy boolee misaal yu yàgg yiñ denc ci diiru tàggat bu bees
5. Lan la LoRA di def ci diisaayu modelu base suñu yokke benn adaptatër bu jëm ci liggéey bi?
   • It freezes base weights and trains a low-rank adapter, which may reduce direct interference but does not guarantee retention
   • It rewrites every base weight for each new task
   • It stores no task-specific parameters
   • It guarantees zero forgetting on every task sequence
6. naka lay def ba gis-gis bu yokk ci recuperation moytu fàtte bu metti?
   • Ni ñuy efaase xam-xam bu yàgg ngir am palaas
   • Sooy denc xam-xam bu bees ci benn base de done ci biti ci barabu poid yi
   • Sooy tàggat sa bopp ci bruit yu bari
   • Ni ñuy dindi mémoire model bi yépp
7. How does Gradient Episodic Memory (GEM) constrain learning on a new task?
   • It removes old task examples from memory
   • It changes the model to a nearest-class-mean classifier
   • It adjusts the gradient so the first-order approximation does not increase loss on stored prior-task examples
   • It adds a KL penalty to every parameter regardless of task
8. How do RAG-Token and RAG-Sequence differ in the original RAG paper?
   • RAG-Token removes retrieval while RAG-Sequence removes generation
   • RAG-Sequence uses images while RAG-Token uses audio
   • RAG-Token trains without model parameters
   • RAG-Token marginalizes retrieved documents per output token, while RAG-Sequence uses one document for a whole sequence