[2605.31183] Steering LLMs? Actually, Sparse Autoencoders can outperform simple baselines
![[2605.31183] Steering LLMs? Actually, Sparse Autoencoders can outperform simple baselines](/static/img/curated/afbytes-us.png)
AI disclosure
Summary
Abstract page for arXiv paper 2605.31183: Steering LLMs? Actually, Sparse Autoencoders can outperform simple baselines