Sarraynta iyo Kediska Badan ee Semanticity

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Waa maxay 'superposition' waxa loola jeedaa shabakadaha neerfaha?
   • Matalaya sifooyin badan oo ka badan kuwa neurons-yada iyadoo la isku dhejinayo jihooyinkooda
   • Isku dhufashada noocyo badan oo midba midka kale ah
   • Ku socodsiinta moodalka laba jeer iyo celcelis ahaan
   • Ku darida buuqa agabka inta lagu jiro tababarka
2. Xaalad noocee ah ayaa ka dhigaysa kor-u-qaadista inay si fiican u shaqeyso inkasta oo faragelin muuqaal ah?
   • Astaamaha oo si cufan u firfircoon mar kasta
   • Astaamaha oo aad u yar, sidaa darteed in yar ayaa isku mar wada firfircoon
   • Isticmaalka firfircoonida wanaagsan oo kaliya
   • Lahaanshaha neurons ka badan sifooyinka
3. Waa maxay neuron 'polysemantic'?
   • Neuron aan waligii shaqayn
   • Neuron loo isticmaalo kaliya lakabka wax soo saarka
   • Neuron ka jawaabaya fikrado badan oo aan xiriir la lahayn
   • Neuron aan lahayn wax xiriir ah
4. Waa kuwee Anthropic warqad soo bandhigtay daraasad la xakameeyey oo ku saabsan dhacdadan 2022?
   • Feejignaan Waa Dhammaan Waxaad U Baahan Tahay
   • Kala soocida ImageNet ee leh CNN-yada qoto dheer
   • Ku ciyaarista Atari Deep RL
   • Moodooyinka Ciyaartoyga ee Sarraynta
5. Haddii lakabku uu kaydiyo sifooyin badan oo ka badan inta uu le'eg yahay, maxaa joomatari ahaan lama huraan ah?
   • Astaamaha qaar waa inay wadaagaan jihooyin is dulsaaran, oo aan toos ahayn
   • Sifooyinku dhammaantood waxay ahaan karaan kuwo si qumman u ahaan kara orthogonal
   • Shabakadu waa inay burburtaa
   • Dhammaan sifooyinka waxay noqonayaan isku mid
6. Waa maxay sababta sare-u-qaadista ay u adkeynayso fasiraadda?
   • Waxay ka dhigaysa moodelku inuu si gaabis ah u ordo
   • Neurons-ka kelida ahi si nadiif ah uma maareeyaan fikrado keli ah
   • Waxay ka saartaa dhammaan sifooyinka qaabka
   • Waxa kaliya oo ay saamaysaa lakabka gelinta
7. Why does sparse feature activity reduce interference when more features than dimensions are represented?
   • Sparse activity makes all feature directions orthogonal
   • It increases the number of neurons automatically
   • Fewer features are active at the same time, so overlapping directions collide less often
   • It removes the need for nonlinear computation
8. What does an antipodal pair of feature directions look like geometrically?
   • Two vectors at a right angle in separate dimensions
   • Two vectors with identical directions and magnitudes
   • A vector with no coordinates
   • Two directions that point opposite ways along the same line