Diidmada Muunad bixinta Fine-Tuning Kediska

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Which loop best describes how Rejection Sampling Fine-Tuning improves a model?
   • Samee jawaabaha musharraxiinta badan, hay kuwa ugu dhibcaha badan, oo dib u tababar iyaga
   • Tirtir lakabyada moodeelka ugu waxqabadka xun
   • Diid dhammaan xogta tababarka oo dib u bilow
   • Tusaalaha hal mar oo had iyo jeer aqbal wax soo saarka
2. Goobaha la xaqiijin karo sida xisaabta ama koodka, maxaa u adeegi kara abaal-marin ahaan?
   • Qaabka abaalgudka la bartay oo keliya
   • Hadday jawaabtu sax tahay ama imtixaanadu ha dhaafaan
   • Dhererka jawaabta
   • Tirada calaamadaha la sameeyay
3. Which model family does the cited paper describe using rejection sampling in its alignment pipeline?
   • GPT-4o
   • BERT Base
   • Meta’s Llama 2-Chat
   • Whisper Large
4. Why might a team choose rejection-sampling fine-tuning over a policy-gradient approach for some tasks?
   • It never needs a candidate generator or scoring rule
   • It guarantees no reward hacking
   • It always requires less compute than every RL method
   • It can use supervised learning on selected generations instead of updating through policy-gradient RL
5. Tababarka ku saabsan muunadaha abaal-marinta ugu sarreeya si wax ku ool ah muxuu sameeyaa?
   • Waxay u kala saartaa tayada ugu wanaagsan ee N soo saarida hal muunada ee moodeelka
   • Waxay kordhisaa cabbirka erayada
   • Waxa ay meesha ka saartaa awooda moodeelka in uu qoraal soo saaro
   • Laba jibaaraya tirinta cabbirka
6. What should teams monitor when repeatedly fine-tuning on filtered model-generated outputs?
   • Only the number of model parameters
   • Output diversity alongside held-out correctness
   • Whether every answer uses the same wording
   • Only the training loss on selected samples
7. What is a limitation of rejection-sampling fine-tuning’s candidate selection?
   • It can only be used with image labels
   • It always changes the tokenizer
   • It cannot train on a high-quality answer that was never generated or retained
   • It makes the scoring function unnecessary
8. What risk arises if the scoring model used to filter candidates is flawed?
   • The policy becomes unable to generate any text
   • The tokenizer automatically changes vocabulary
   • The reference labels become unnecessary
   • The policy may learn to exploit the scorer rather than improve the intended quality