Kediska Xisaabinta-Waqtiga Imtixaanka

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Waa maxay ' xisaabinta-waqtiga-tijaabada' loola jeedaa?
   • Xisaabinta waxay ku bixisay tababarida moodalka xogta
   • Wakhtiga ay qaadanayso soo dejinta model
   • Xisaabinta ayaa la kharash gareeyaa marka moodeelku ka jawaabo su'aal la tixraacayo
   • Korontada loo isticmaalo in lagu qaboojiyo xarumaha xogta
2. In test-time scaling studies, what can extra inference compute be used for?
   • Changing the model’s training data during the request
   • Increasing the number of model parameters automatically
   • Guaranteeing correctness without evaluation
   • Searching or revising candidate solutions and checking their quality
3. Waa maxay ganacsiga ubucda u ah xisaabinta wakhtiga-tijaabada?
   • Saxnaanta sare ee beddelka daahitaanka iyo kharash badan
   • Xusuus yar laakiin tababarka gaabis ah
   • Xog badan laakiin moodooyinka yaryar
   • Sawirro ka wanaagsan xallinta hoose
4. Waa maxay muunada 'ugu-wanaagsan-N'?
   • Tababarka N ugu weyn ee xogta
   • Soo saarista jawaabaha musharraxiinta N iyo doorashada midka ugu fiican
   • Isticmaalka kaliya ereyada N ugu horreeya ee degdegga ah
   • Ku socodsiinta moodelka on kombiyuutarada kala duwan N ee xawaaraha
5. Waa maxay sababta xisaabinta-waqtiga-tijaabada loogu tixgaliyo ' dhidibka is-qabashada' cusub?
   • Waxay bedelaysaa baahida loo qabo tababar kasta
   • Waxay kaliya ku shaqeysaa moodooyinka yaryar
   • Waxay wanaajisaa waxqabadka iyadoo ku daraysa xisaabinta fikradda halkii ay ka ahaan lahayd koraya cabbirka moodeelka
   • Waxay yaraynaysaa tirada isticmaalayaasha moodelku u adeegi karo
6. Waa maxay natiijada la yaabka leh ee xisaabinta wakhtiga imtixaanku awood u leedahay?
   • Moodooyinka waxay socon karaan iyada oo aan wax GPUs ah
   • Xogta tababarku waxay noqotaa mid aan loo baahnayn
   • Moodooyinka had iyo jeer waxay u jawaabaan si degdeg ah
   • Qaabka yar ee fikirka dheer wuxuu la tartami karaa moodal aad u weyn oo isla markiiba u jawaabaya
7. What is one reason a fixed best-of-N strategy may waste inference compute?
   • It spends the same candidate budget on easy and difficult prompts
   • It modifies model weights on every query
   • It cannot generate multiple candidates
   • It eliminates the need for scoring
8. What did the test-time compute paper compare with best-of-N sampling?
   • A strategy that disables inference
   • A compute-optimal strategy that allocates effort based on expected solution quality
   • A larger tokenizer vocabulary
   • A method that changes image resolution