Derives optimality constraints for nonnegative joint dictionary learning that explain observed SAE behaviors such as feature splitting, absorption, and dense antipodal features.
arXiv preprint arXiv:1908.09961 , year=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
q-bio.NC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
How Optimality Structures Sparse Dictionaries: A Theory for Understanding SAE Representations
Derives optimality constraints for nonnegative joint dictionary learning that explain observed SAE behaviors such as feature splitting, absorption, and dense antipodal features.