Maswali yanayounganishwa na mwongozo · Kati Kiwango

Maswali ya Uboreshaji ya Kahneman-Tversky

Hukagua maoni ya kuhitajika kwa binary ya KTO, motisha ya nadharia-tarajiwa, zawadi za pointi za marejeleo, na mipaka katika mipangilio yote.

Njia za mwongozo zinazohusianaUboreshaji wa Kahneman Tversky
Swali 1 ya 8

KTO inahitaji data ya aina gani?

Hakiki kila swali8 maswali

Majibu na maelezo huonekana unapojibu maswali.

  1. KTO inahitaji data ya aina gani?
    • Lebo za binary zinazohitajika/zisizohitajika kwenye matokeo ya mtu binafsi
    • Jozi zilizoorodheshwa za majibu yaliyochaguliwa na kukataliwa
    • Maandishi yasiyo na lebo pekee
    • Alama za nambari za zawadi kutoka 0 hadi 100
  2. KTO imehamasishwa na kazi gani?
    • Uhusiano wa Einstein
    • Nadharia ya matarajio ya Kahneman na Tversky
    • Maendeleo ya Darwin
    • Nadharia ya habari ya Shannon
  3. Je, 'chukizo la hasara' kama linavyotumika katika KTO ni nini?
    • Epuka utendakazi wowote wa upotezaji kabisa
    • Kukataa kutoa mafunzo juu ya hasara
    • Kuadhibu matokeo yasiyofaa kwa ukali zaidi kuliko kuthawabisha yale yanayohitajika
    • Kufanya kiwango cha kujifunza kuwa hasi
  4. What kind of feedback can KTO learn from without requiring paired comparisons?
    • Only a ranked list of five outputs
    • Only a reference translation
    • Only a token-level grammar mask
    • A binary label indicating whether an individual output is desirable
  5. Katika KTO, kupotoka hupimwa kulingana na nini?
    • Sehemu ya marejeleo au kiwango cha msingi cha zawadi
    • Jumla ya idadi ya vigezo
    • Kiwango cha joto cha GPU
    • Saizi ya msamiati
  6. What does the KTO reference point help define?
    • The number of transformer layers
    • Whether an output’s implicit reward is above or below a baseline
    • A fixed translation in another language
    • The model’s maximum context length
  7. Does KTO’s use of binary desirability feedback prove it will outperform preference-pair methods?
    • Yes, binary feedback is always superior
    • Yes, it eliminates the need to train a model
    • No; performance depends on the model, data, and evaluation setting
    • No, because binary feedback cannot be used for language models
  8. Which behavioral-economics framework motivates KTO?
    • Classical signal detection theory
    • Bayesian network structure learning
    • The efficient-market hypothesis
    • Kahneman and Tversky’s prospect theory