WriteSAE is the first sparse autoencoder that factors decoder atoms into the native d_k x d_v cache write shape of recurrent models and supplies a closed-form per-token logit shift for atom substitution.
arXiv preprint arXiv:2504.13151 , year=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
Gender bias and factual gender knowledge are severely entangled in language model circuits and neurons, making neuron ablation an unreliable method for debiasing.
citing papers explorer
-
WriteSAE: Sparse Autoencoders for Recurrent State
WriteSAE is the first sparse autoencoder that factors decoder atoms into the native d_k x d_v cache write shape of recurrent models and supplies a closed-form per-token logit shift for atom substitution.
-
GKnow: Measuring the Entanglement of Gender Bias and Factual Gender
Gender bias and factual gender knowledge are severely entangled in language model circuits and neurons, making neuron ablation an unreliable method for debiasing.