Maswali ya Kuongeza Kuongeza Muda wa Mtihani

Majibu na maelezo huonekana unapojibu maswali.

1. Je, 'kuhesabu muda wa majaribio' inarejelea nini?
   • Uhesabuji uliotumika kufunza modeli kwenye data
   • Wakati inachukua kupakua mfano
   • Uhesabuji uliotumika wakati modeli inajibu swali kwa uelekezaji
   • Umeme unaotumika kupoza vituo vya data
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. Je, ni biashara gani kuu ya kuongeza muda wa majaribio?
   • Usahihi wa juu zaidi kwa kubadilishana kwa muda na gharama zaidi
   • Kumbukumbu kidogo lakini mafunzo ya polepole
   • Data zaidi lakini mifano ndogo
   • Graphics bora kwa azimio la chini
4. Sampuli ya 'bora-ya-N' ni nini?
   • Mafunzo kwenye hifadhidata kubwa zaidi za N
   • Inazalisha majibu ya mtahiniwa wa N na kuchagua lililo bora zaidi
   • Kwa kutumia maneno ya N ya kwanza pekee ya kidokezo
   • Kuendesha modeli kwenye kompyuta N tofauti kwa kasi
5. Kwa nini hesabu ya muda wa majaribio inachukuliwa kuwa 'mhimili mpya wa kuongeza alama'?
   • Inachukua nafasi ya hitaji la mafunzo yoyote
   • Inafanya kazi tu kwa mifano ndogo
   • Inaboresha utendakazi kwa kuongeza hesabu ya uelekezaji badala ya kukuza ukubwa wa kielelezo pekee
   • Inapunguza idadi ya watumiaji ambao mtindo unaweza kutumika
6. Ni matokeo gani ya kushangaza ambayo hesabu ya wakati wa jaribio inawasha?
   • Miundo inaweza kukimbia bila GPU yoyote
   • Data ya mafunzo inakuwa si lazima
   • Mifano daima hujibu haraka
   • Mfano mdogo unaofikiria kwa muda mrefu unaweza kushindana na mtindo mkubwa zaidi unaojibu mara moja
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