Aligned training reparameterizes SAEs to enforce unit inner product between encoder and decoder directions, eliminating dead features and enhancing stability without hyperparameters.
Enhancing neural network interpretability with feature-aligned sparse autoencoders
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
years
2026 3verdicts
UNVERDICTED 3representative citing papers
A cross-attention SAE with sparsemax attention achieves lower reconstruction loss and higher-quality concepts than fixed-sparsity baselines by making activation counts data-dependent.
Sparse autoencoders enable phase synchronization in frozen graph CFD surrogates through Hilbert-identified oscillatory features and SVD-based time-varying rotations.
citing papers explorer
-
Aligned Training: A Parameter-Free Method to Improve Feature Quality and Stability of Sparse Autoencoders (SAE)
Aligned training reparameterizes SAEs to enforce unit inner product between encoder and decoder directions, eliminating dead features and enhancing stability without hyperparameters.
-
Improving Sparse Autoencoder with Dynamic Attention
A cross-attention SAE with sparsemax attention achieves lower reconstruction loss and higher-quality concepts than fixed-sparsity baselines by making activation counts data-dependent.
-
Sparse Autoencoders as a Steering Basis for Phase Synchronization in Graph-Based CFD Surrogates
Sparse autoencoders enable phase synchronization in frozen graph CFD surrogates through Hilbert-identified oscillatory features and SVD-based time-varying rotations.