Adalu ti awọn ijinle adanwo

Awọn idahun ati awọn alaye han bi o ṣe n mu ibeere naa.

1. Kini Adalu awọn ijinle yatọ kọja awọn ami?
   • Elo ni isiro (eyi ti fẹlẹfẹlẹ) kọọkan àmi gba
   • Awọn fokabulari ti a lo
   • Awọn data ikẹkọ
   • Èdè àbájáde
2. Bawo ni MoD ṣe jẹ asọtẹlẹ isuna iṣiro rẹ?
   • Nipa sisọ idaji awọn iwọn awoṣe silẹ laileto
   • Nipa sisẹ ida oke-k ti o wa titi ti awọn ami ni bulọọki kọọkan
   • Nipa lilo awoṣe ti o yatọ fun gbolohun kọọkan
   • Nipa pipa akiyesi patapata
3. Kini yoo ṣẹlẹ si awọn ami ti olulana KO yan ni bulọọki ti a fun?
   • Wọn ti paarẹ lati ọkọọkan
   • Wọn ti wa ni rán si lọtọ awoṣe
   • Wọn foju bulọki nipasẹ asopọ ti o ku, ti o kọja nipasẹ ko yipada
   • Wọn ti wa ni iyipada si awọn aworan
4. Which lab published the Mixture of Depths technique?
   • OpenAI ni ọdun 2019
   • Meta ni ọdun 2015
   • Microsoft ni ọdun 2012
   • Google DeepMind ni ọdun 2024
5. Kini apapọ MoD pẹlu Adapọ-ti-Amoye gbejade?
   • Ipo, eyiti awọn ipa-ọna lori ijinle mejeeji ati yiyan iwé
   • Awoṣe pẹlu ko si paramita
   • A funfun convolutional nẹtiwọki
   • A tokenizer-nikan eto
6. Kini idi ti oke-k ti o wa titi ṣe fẹ ju iloro-ami-ami ni MoD?
   • O mu ki awọn awoṣe o wu ID ọrọ
   • O ntọju tensor ni nitobi aimi, eyi ti o jẹ daradara lori hardware
   • O yọ iwulo fun ikẹkọ kuro
   • O mu ki awọn fokabulari iwọn
7. What does the learned router do at a Mixture-of-Depths block?
   • Selects a new model family for every prompt
   • Predicts labels for the next training batch
   • Scores tokens so the block can select which ones receive full attention and MLP computation
   • Changes the vocabulary size at inference
8. Why does causal generation in MoD use auxiliary predictors for routing?
   • To let a model read the entire answer before generating
   • To remove the fixed top-k compute budget
   • To convert the sequence into an image
   • To make routing decisions without looking at future tokens