Nukleus iyo Top-k Kediska Muunada

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Sidee buu muunada top-k u xaddidaa doorashooyinka calaamada?
   • Waxay ilaalinaysaa oo keliya calaamadaha ixtimaalka ugu sarreeya, ka dibna muunado iyaga ka mid ah
   • Waxay haysaa calaamado ilaa ay ixtimaalka isugaynta ahi dhaafto k
   • Had iyo jeer waxay doorataa calaamadda ugu macquulsan
   • Waxay meesha ka saartaa calaamadaha ugu yar ee kaliya
2. Maxaa qeexaya 'nucleus' ee ku jira muunada nukleus (top-p)?
   • Erayada k ugu badan ee xogta tababarka
   • Tusmada ugu yar ee ixtimaalkahoodu ay dhan yihiin ugu yaraan p
   • Calaamadda ixtimaalka ugu sarreeya ee keliya
   • Dhammaan calaamadaha leh itimaalka ka sarreeya 50%
3. Waa maxay faa'iidada muhiimka ah ee top-p ka sarreeya top-k?
   • Had iyo jeer way dhakhso badan tahay in la xisaabiyo
   • Marna ma soo saarto wax soo saar random
   • Tirada musharraxiintu waxay la qabsanayaan kalsoonida moodeelka
   • Waxay iska indha tirtaa itimaalka calaamadda
4. Muxuu sameeyaa cabbirka heerkulku ka hor tijaabada?
   • Waxay dejisaa dhererka isku xigxiga ee ugu sarreeya
   • Waxay doorataa luqadda wax-soo-saarka
   • Waxay tirisaa cabbirka erayada
   • Waxay dib u miisaamaysaa qaybinta ixtimaalka inay noqoto mid fiiqan ama sasabasan
5. What result did the nucleus-sampling paper report for its tested open-ended generation tasks?
   • It improved diversity without sacrificing fluency and coherence in the reported comparisons
   • It guaranteed factual correctness in every output
   • It eliminated all sampling randomness
   • It improved translation quality on every benchmark
6. For a greedy decoding baseline with no token sampling, how should generation be configured?
   • Disable sampling and use greedy decoding to select the highest-probability token at each step
   • Use high temperature and a large nucleus threshold
   • Randomize the temperature at every step
   • Keep sampling enabled but increase top-p
7. After top-k truncation removes tokens, what happens to the remaining sampling probabilities?
   • They remain assigned their original total mass below one
   • They are replaced with equal probabilities
   • They are renormalized over the retained top-k tokens
   • They are converted into beam-search scores
8. What did Holtzman et al. report about likelihood-maximizing decoding for open-ended text?
   • It always yields the most diverse human-like passages
   • Waxay meesha ka saaraysaa baahida loo qabo qaab luqadeed
   • It guarantees factual correctness
   • It can produce bland and repetitive text despite strong likelihood scores