Maswali yanayounganishwa na mwongozo · Kati Kiwango

Maswali ya EleutherAI

Kagua miundo ya lugha huria ya EleutherAI, seti ya data ya Rundo, zana za kutathmini, na chaguo za usanifu wa vitendo kama vile RoPE.

Njia za mwongozo zinazohusianaEleutherai
Swali 1 ya 8

How did EleutherAI first bring together its research community?

Hakiki kila swali8 maswali

Majibu na maelezo huonekana unapojibu maswali.

  1. How did EleutherAI first bring together its research community?
    • As a Discord server for discussing GPT-3 and open language modeling
    • Kama uanzishaji unaofadhiliwa na mradi huko Silicon Valley
    • Kama maabara ya utafiti ya serikali
    • As a university’s official AI department
  2. 'The Pile' ni nini?
    • Bidhaa ya chatbot inayomilikiwa
    • Seti ya data iliyoratibiwa ya maandishi iliyo wazi ya GB 825 kwa miundo ya lugha ya mafunzo
    • Aina ya safu ya mtandao wa neva
    • Jukwaa la kompyuta la wingu la EleutherAI
  3. Ni ipi kati ya hizi ni modeli ya lugha ya EleutherAI?
    • BERT-Kubwa
    • LLaMA-2
    • GPT-NeoX-20B
    • PALM
  4. Kiunga cha Tathmini cha LM kinatumika kwa ajili gani?
    • Inazalisha picha kutoka kwa maandishi
    • Kukandamiza uzito wa mfano
    • Miundo ya upangishaji kama API za wavuti
    • Kulinganisha miundo ya lugha katika kazi nyingi
  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 mabadiliko gani ya shirika ambayo EleutherAI ilifanya mnamo 2023?
    • Ilinunuliwa na OpenAI
    • Ilirasimishwa kama taasisi ya utafiti isiyo ya faida
    • Ikawa mfuko wa ua wa faida
    • Ilizima kabisa
  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