Strength Controls Change How Interventions Compare Across Sparse Autoencoder Codes
Changing sparse autoencoder (SAE) codes can alter both intervention direction and magnitude, so comparing intervention effects requires a strength control specifying what is held equal.
ProofPaper ↗
Key points
- Norm matching and output calibration provide different controls, but whether interventions yield different effects across codes after strength matching remains unclear.
- We compare intervention effects under these controls for fixed feature groups in Gemma and Llama and observe: (1) Whole-patch scaling.
- With one coefficient per input, full-output Kullback-Leibler (KL) calibration leaves smaller mean effect gaps from the encoder-based reference intervention than total-norm matching for nonnegative least-squares (NNLS) refits.
- On a small matched gender-agreement subset, positionwise norm matching leaves smaller observed mean gaps than full-output calibration.
Sources (1)
- [1]Strength Controls Change How Interventions Compare Across Sparse Autoencoder CodesarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 19, 02:29 PM
Changing sparse autoencoder (SAE) codes can alter both intervention direction and magnitude, so comparing intervention effects requires a strength control specifying what is held equal.
Norm matching and output calibration provide different controls, but whether interventions yield different effects across codes after strength matching remains unclear.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 19, 2026EvalRx: Agentic Diagnosis and Verified Repair of Foundation Models
- Oct 19, 2026From Black Box to Controller: Inference-Time Steering of LLM via Feature Amplification
- Oct 18, 2026CHSM8: A Chess Transformer that Follows Natural-Language Style Prompts
- Oct 17, 2026A Symmetry-Matching Approach to Blind-Spot Reduction in Sparse Autoencoders
- Oct 17, 2026llama.cpp has come a long way
- Oct 16, 2026Beyond SAEs: Consistent Concept Discovery with Sparse Coding