Idanwo ti o sopọ mọ itọsọna · Alabọde Ipele

Yi Models nipasẹ 01.AI adanwo

Review the Yi model family’s origins, bilingual focus, decoder-only design, long-context variants, and instruction-oriented post-training.

Jẹmọ itọsọna awọn ọnaYi Awọn awoṣe Nipa 01 AI
Ibeere 1 ti 8

Tani o da 01.AI, ile-iṣẹ lẹhin awọn awoṣe Yi?

Awotẹlẹ gbogbo ibeere8 ibeere

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

  1. Tani o da 01.AI, ile-iṣẹ lẹhin awọn awoṣe Yi?
    • Kai-Fu Lee
    • Robin Li
    • Jack Ma
    • Andrew Ng
  2. Kini meji ti awọn awoṣe ipilẹ ipilẹ Yi akọkọ ti tu silẹ?
    • Yi-1 ati Yi-2
    • Yi-6B og Yi-34B
    • Yi-100 ati Yi-200
    • Yi-Mini og Yi-Max
  3. Agbara wo ni awọn awoṣe Yi ti a mọ ni pataki fun?
    • Ṣiṣẹda awọn aworan nikan
    • Nṣiṣẹ laisi GPU eyikeyi
    • Iṣẹ ṣiṣe ti o lagbara ni ede meji-ede Kannada-English
    • Ohùn sisọ nikan ni itumọ
  4. Iru iran faaji ipilẹ wo ni awọn awoṣe Yi tẹle?
    • Apẹrẹ LSTM loorekoore
    • A awoṣe itankale
    • Awoṣe kooduopo mimọ bi BERT
    • Oluyipada koodu-nikan ti ara Llama
  5. Ni aijọju bawo ni ferese ọrọ ọrọ ti awọn iyatọ Yi ti o gun-gun ṣe pẹ to?
    • 200K àmi
    • 4K àmi
    • 32K àmi
    • 1 million àmi
  6. What did 01.AI’s Yi technical report identify as an important driver of model performance?
    • Kọ AI silẹ patapata
    • Data quality resulting from data-engineering efforts
    • Nikan ile hardware awọn eerun
    • Ṣepọ pẹlu OpenAI
  7. What does supervised fine-tuning plus reinforcement learning from human feedback aim to improve in Yi chat models?
    • Compressing retinal images
    • Changing the model into an encoder-only network
    • Following instructions in conversational use
    • Removing all text-generation ability
  8. What data strategy did 01.AI emphasize for the Yi models?
    • Training exclusively on synthetic audio
    • Avoiding all data filtering
    • Using only programming-language source code
    • Careful data quality and curation rather than relying only on sheer scale