Kedis ku xidhan hagaha · Dhexdhexaad Heerka

EleutherAI Kediska

Dib u eeg moodooyinka luqadda furan ee EleutherAI, kaydka xogta, agabka qiimaynta, iyo doorashooyinka qaab dhismeedka la taaban karo sida RoPE.

Wadooyinka haga ee la xidhiidhaEleutheri
Su'aal 1 ee 8

How did EleutherAI first bring together its research community?

Horudhac su'aal kasta8 su'aalo

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

  1. How did EleutherAI first bring together its research community?
    • As a Discord server for discussing GPT-3 and open language modeling
    • Sida bilawga maalgelinta maalgelinta ee Silicon Valley
    • Sida shaybaadhka cilmi-baarista dawladda
    • As a university’s official AI department
  2. Waa maxay 'Pile'?
    • Alaab chatbot iska leh
    • 825GB xog-qoraal oo la diyaariyey oo furan oo loogu talagalay qaababka luqadda
    • Nooc ka mid ah lakabka shabakada neerfaha
    • Madal xisaabinta daruuraha ee EleutherAI
  3. Waa kuwee kuwan nooca luqadda EleutherAI?
    • BERT-weyn
    • LLMA-2
    • GPT-NeoX-20B
    • PaLM
  4. Waa maxay suunka Qiimaynta LM ee loo isticmaalo?
    • Ka soo saarida sawirada qoraalka
    • Cadaadiska miisaanka moodada
    • Qaababka martigelinta sida API-yada webka
    • Habbeegsiyeedka qaababka luqadda ee hawlo badan
  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. Waa maxay isbeddelka urureed ee EleutherAI ay samaysay 2023?
    • Waxaa la wareegay OpenAI
    • Waxa loo qaabeeyey sidii machad cilmi-baadhiseed oo aan faa'iido doon ahayn
    • Waxa ay noqotay sanduuq macaash doon ah
    • Gebi ahaanba waa la xidhay
  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