arXiv Machine Learning#tech
From Sparse Features to Trustworthy Proxies: Certifying SAE-Based Interpretabilitytranslating…
Factuality: 61/100UnknownarXiv Digital Media
arXiv:2606.18383v1 Announce Type: new
Abstract: Sparse autoencoders (SAEs) are increasingly used to extract interpretable features from language models (LMs), yet a central question remains: when can an SAE-based explanation be treated as a faithful view of an underlying frozen LM We study this through a post-hoc generalization framework that certifies the LM via a sparse proxy, obtained by replacing a native hidden activation with its pretrained SAE reconstruction. Our framework derives an upp