pith. sign in

Towards monoseman- ticity: Decomposing language models with dictionary learn- ing.Transformer Circuits Thread, 2023

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

fields

cs.LG 1

years

2026 1

verdicts

UNVERDICTED 1

representative citing papers

Improving Sparse Autoencoder with Dynamic Attention

cs.LG · 2026-04-16 · unverdicted · novelty 7.0

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.

citing papers explorer

Showing 1 of 1 citing paper.

  • Improving Sparse Autoencoder with Dynamic Attention cs.LG · 2026-04-16 · unverdicted · none · ref 4

    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.