all AI news
Polysemanticity and Capacity in Neural Networks. (arXiv:2210.01892v1 [cs.NE])
Oct. 6, 2022, 1:11 a.m. | Adam Scherlis, Kshitij Sachan, Adam S. Jermyn, Joe Benton, Buck Shlegeris
cs.LG updates on arXiv.org arxiv.org
Individual neurons in neural networks often represent a mixture of unrelated
features. This phenomenon, called polysemanticity, can make interpreting neural
networks more difficult and so we aim to understand its causes. We propose
doing so through the lens of feature \emph{capacity}, which is the fractional
dimension each feature consumes in the embedding space. We show that in a toy
model the optimal capacity allocation tends to monosemantically represent the
most important features, polysemantically represent less important features (in
proportion to …
More from arxiv.org / cs.LG updates on arXiv.org
Jobs in AI, ML, Big Data
(373) Applications Manager – Business Intelligence - BSTD
@ South African Reserve Bank | South Africa
Data Engineer Talend (confirmé/sénior) - H/F - CDI
@ Talan | Paris, France
Data Science Intern (Summer) / Stagiaire en données (été)
@ BetterSleep | Montreal, Quebec, Canada
Director - Master Data Management (REMOTE)
@ Wesco | Pittsburgh, PA, United States
Architect Systems BigData REF2649A
@ Deutsche Telekom IT Solutions | Budapest, Hungary
Data Product Coordinator
@ Nestlé | São Paulo, São Paulo, BR, 04730-000