LESSWRONGTags
LW

Sparse Autoencoders (SAEs)

•

Applied to Quick Thoughts on Scaling Monosemanticity by Joel Burget 9d ago

•

Applied to SAE sparse feature graph using only residual layers by crayhippo 9d ago

•

Applied to Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning by Dan Braun 15d ago

•

Applied to Towards Multimodal Interpretability: Learning Sparse Interpretable Features in Vision Transformers by Joseph Bloom 1mo ago

•

Applied to Transcoders enable fine-grained interpretable circuit analysis for language models by Jacob Dunefsky 1mo ago

•

Applied to Improving Dictionary Learning with Gated Sparse Autoencoders by Neel Nanda 1mo ago

•

Applied to ProLU: A Nonlinearity for Sparse Autoencoders by Noa Nabeshima 1mo ago

•

Applied to Discriminating Behaviorally Identical Classifiers: a model problem for applying interpretability to scalable oversight by Neel Nanda 1mo ago

•

Applied to Past Tense Features by Can 1mo ago

•

Applied to [Summary] Progress Update #1 from the GDM Mech Interp Team by Joseph Bloom 1mo ago

•

Applied to [Full Post] Progress Update #1 from the GDM Mech Interp Team by Joseph Bloom 1mo ago

•

Applied to Experiments with an alternative method to promote sparsity in sparse autoencoders by Eoin Farrell 2mo ago

•

Applied to Normalizing Sparse Autoencoders by Fengyuan Hu 2mo ago

•

Applied to Improving SAE's by Sqrt()-ing L1 & Removing Lowest Activating Features by Joseph Bloom 2mo ago

•

Applied to SAE-VIS: Announcement Post by Joseph Bloom 2mo ago

•

Applied to SAE reconstruction errors are (empirically) pathological by Joseph Bloom 2mo ago

•

Applied to My best guess at the important tricks for training 1L SAEs by Joseph Bloom 2mo ago

•

Applied to Addressing Feature Suppression in SAEs by Joseph Bloom 2mo ago