Ikibazo gihuza ibibazo · Hagati Urwego

Ikibazo cya EleutherAI

Subiramo EleutherAI yerekana imvugo ifunguye, Ikirundo cyamakuru, ibikoresho byo gusuzuma, hamwe nuburyo bwo guhitamo bwubaka nka RoPE.

Inzira zijyanye no kuyoboraEleutherai
Ikibazo 1 Bya 8

How did EleutherAI first bring together its research community?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. How did EleutherAI first bring together its research community?
    • As a Discord server for discussing GPT-3 and open language modeling
    • Nkumushinga uterwa inkunga na rwiyemezamirimo mu kibaya cya Silicon
    • Nka laboratoire ya leta
    • As a university’s official AI department
  2. 'Ikirundo' ni iki?
    • Igicuruzwa cyihariye
    • 825GB ifunguye inyandiko yatunganijwe kugirango yimenyereze imvugo yindimi
    • Ubwoko bw'urusobe rw'imitsi
    • Ikibanza cyo kubara ibicu bya EleutherAI
  3. Ninde muribi murugero rwururimi rwa EleutherAI?
    • BERT-Kinini
    • LLaMA-2
    • GPT-NeoX-20B
    • PaLM
  4. Niki LM Isuzuma Harness ikoreshwa?
    • Gukora amashusho kuva mwandiko
    • Kugabanya uburemere bw'icyitegererezo
    • Kwakira moderi nkurubuga APIs
    • Ibipimo byerekana ururimi mubikorwa byinshi
  5. Which organizations are named in EleutherAI’s retrospective as research supporters?
    • CoreWeave and Google TPU Research Cloud
    • Nvidia and the National Science Foundation only
    • The U.S. Department of Energy and NASA
    • Amazon and the European Space Agency
  6. Ni irihe hinduka ry'umuteguro EleutherAI yakoze muri 2023?
    • Yaguzwe na OpenAI
    • Yashyizwe mubikorwa nkikigo cyubushakashatsi kidaharanira inyungu
    • Yabaye ikigega cyunguka inyungu
    • Yahagaritse burundu
  7. Which positional method does the GPT-NeoX-20B paper describe using instead of learned absolute position embeddings?
    • A separate learned embedding for every document
    • A convolutional image-position map
    • Rotary Positional Embeddings (RoPE)
    • A random position permutation at each layer
  8. What architectural arrangement is highlighted as a GPT-NeoX-20B difference from GPT-3?
    • The model replaces its Transformer with a decision tree
    • Every layer removes the attention mechanism
    • It uses only an encoder and no autoregressive decoder
    • Attention and feed-forward layers can be computed in parallel