Quiz buy tekki lu mekanik

Sooy def quiz bi, tontu yi ak leeral yi dañuy feeñ.

1. Lan mooy mébetu tekkib mekanik?
   • Ngir soppi xayma yi ci biir reso bi ci algorithm yu nit ñi mëna xam
   • Ngir model yi gëna gaaw ci tàggat yaram
   • Ngir kompresse ay model ngir aparey mobile
   • Ngir rang màndarga yiñ dugal ci soloom kese
2. Luy 'superposition' ci biir reso neuronal?
   • boole ay model yu bari
   • Jëfandikoo yoon yuy jaxasoo di wane màndarga yu bari yu ëpp dimension yi
   • Doxal benn model ci GPU yu bari
   • Liggéeyukaay yi bawoo ci ay daw yu wuute
3. Autoencoder yu néew yi ñu ngi leen di jëfandikoo ci tekki mekanik, li ci njëkk mooy...
   • Defar ay done tàggat
   • Gaawaay ci gis-gis
   • Taxawal jëfandikoo polysemantic ci anam yu gëna set, gëna mëna tekki
   • Encrypter poids model
4. Luy 'boppu induksioŋ'?
   • gaawaayu aparey ngir trafo
   • Layer yiy génne mbind
   • Benn xeetu fonction perte
   • Bopp yiy bàyyi xel ci mbir yiy kopie ak wéyal motif token yu bari, jàppale jàng ci biir contexte
5. Nan la gëstukat yi di mëna firndee ni sircuit biñ gis dafay def dëgg-dëgg xayma biñ xalaat?
   • Jaaraleko ci ay paj yu melni ablating wala patching ay aktivasioŋ yuñ tànn
   • Sooy saytu njubteg tàggat model bi
   • Ni ñuy lim limu parametre yi
   • Ni ñuy nattee latency
6. What is a cautious safety-related reason to study mechanistic interpretability?
   • It proves every hidden behavior can already be found
   • It may provide evidence about internal computations, while current methods are not a complete audit or direct steering guarantee
   • It guarantees a model will follow any requested correction
   • It replaces evaluations of model behavior in deployment
7. If a sparse autoencoder feature activates on examples described as “Golden Gate Bridge,” what does that observation establish?
   • The entire model’s computation is now explained
   • The feature is guaranteed to be a single neuron
   • A candidate feature correlates with those examples and needs further testing to establish its role
   • The feature causes every bridge-related output
8. Why test a candidate feature across varied prompts and examples?
   • Proof that the model has no other relevant circuits
   • A guarantee that the explanation generalizes to every prompt
   • A replacement for inspecting model behavior
   • To check whether the observed activation pattern generalizes beyond selected examples