AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

On the Relationship Between Activation Outliers and Feature Death in Sparse Autoencoders

Elana Simon, Etowah Adams, James Zou

ICML 2026 · Architectures and efficiency › RNNs and normalization · approach: Sparse dictionary representations · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
initialize encoder bias with the activation mean (mean-centering), dict size 2048 (ablation)TopK SAE · ESM3 layer 24 SwissProt concepts (k=16)concepts captured: 61 -> 100Improved
mean-centering, dict size 4096 (ablation)TopK SAE · ESM3 layer 24 SwissProt concepts (k=16)concepts captured: 69 -> 121Improved
mean-centering, dict size 8192 (ablation)TopK SAE · ESM3 layer 24 SwissProt concepts (k=16)concepts captured: 73 -> 127Improved
mean-centering instead of 4x larger dictionaryTopK SAE · ESM3 layer 24 SwissProt concepts (k=16)concepts captured: 73 -> 100Improved
mean-centering on synthetic data (ablation)TopK SAE · Synthetic data (gamma=40)MMCS with ground truth: 0.38 -> 0.97Improved
mean-centering on ESM3 layer 24 dead features (ablation)TopK SAE · ESM3 layer 24 (gamma~8)dead features (%): 75% -> near zeroImproved
add AuxK on ESM3 layer 24 (ablation)TopK SAE · ESM3 layer 24 (gamma~8)dead features (%): 75% -> 25%Improved
freeze the bias during training (ablation)TopK SAE · Synthetic activations (gamma=20)dead-by-ReLU recovery: permanently deadHurt
freeze the bias during training (ablation)TopK SAE · Synthetic activations (gamma=20)dead-by-TopK recoveryNo clear effect
add AuxK loss (ablation)TopK SAE · Synthetic activations (gamma=20)bias convergence speedNo clear effect
mean-centering on Evo1 (extremely low-rank activations) (ablation)TopK SAE · Evo1dead features (%): 73% dead at initialization, ~58% after trainingHurt
apply PCA whitening after centering (ablation)TopK SAE · Evo1 and other low-rank layersdead features (%): 0%Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy