Ikibazo gihuza ibibazo · Hagati Urwego

Igiti cyibitekerezo Ikibazo

Ikizamini cyatekerejweho ibisekuruza, isuzuma rya leta, gushakisha no gusubira inyuma, hiyongereyeho imipaka yihariye yibikorwa byerekana ibisubizo byatanzwe.

Inzira zijyanye no kuyoboraIgiti Cyibitekerezo
Ikibazo 1 Bya 8

Nigute Igiti cyibitekerezo gitandukana cyane nuruhererekane rwibitekerezo?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Nigute Igiti cyibitekerezo gitandukana cyane nuruhererekane rwibitekerezo?
    • Irashakisha amashami menshi yo gutekereza kandi irashobora gusubira inyuma
    • Isubiramo icyitegererezo kumakuru mashya
    • Ikuraho intambwe zose zo hagati zo gutekereza
    • Ikora gusa kubyara amashusho
  2. Nibihe bice bitatu bigize igiti cyibitekerezo?
    • Tokenizer, gushiramo, na decoder
    • Gutanga ibitekerezo, umugenzuzi wa leta, no gushakisha algorithm
    • Encoder, kwitondera umutwe, na softmax
    • Dataset, optimizer, nigikorwa cyo gutakaza
  3. In the original Tree of Thoughts Game of 24 experiment, what result did breadth-five search report for its selected test set?
    • 100% success on every possible puzzle
    • 74% for every model and search budget
    • 74% success for GPT-4 with ToT, versus 4% for a single chain-of-thought baseline
    • 4% for ToT and 74% for chain-of-thought
  4. Nigute sisitemu yigiti cyibitekerezo isuzuma ibisubizo byigice?
    • Kubaza ububikoshingiro byo hanze
    • Mugupima ubushyuhe bwa GPU
    • Kubara umubare wibimenyetso byakoreshejwe
    • Mugusaba icyitegererezo kugereranya leta nkuko byukuri, birashoboka, cyangwa bidashoboka
  5. Ni ubuhe buryo bwo gushyira mu bikorwa Igiti cy'ibitekerezo kuri moderi iriho?
    • Nibisubizo no gushakisha bipfunyika, ntabwo byongeye imyitozo
    • Birasaba kongera imyitozo kuva kera
    • Irashobora gukora gusa kubikoresho byabugenewe
    • Irahindura burundu uburemere bwikitegererezo
  6. Ni ubuhe buryo bufitanye isano butuma inzira zo gutekereza zihuza kandi zigakoresha ibisubizo hagati?
    • Gusubira inyuma
    • Igishushanyo cyibitekerezo
    • Ibisanzwe
    • Gukata ibiti by'ubugari
  7. Which components are commonly specified in a Tree of Thoughts setup?
    • A new tokenizer, an optimizer, and a training corpus
    • A database, a browser, and a GPU driver
    • Thought proposals, an intermediate-state evaluator, and a search procedure
    • A label encoder, a loss scaler, and a checkpoint
  8. Why can Tree of Thoughts increase inference cost?
    • It always retrains the base model
    • It requires a separate model for every token
    • It eliminates all model calls after the first thought
    • It may generate and evaluate multiple candidate states across search steps