Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
Rational Sparse Autoencoder
Naiyu Yin, Yue Yu
Sparse autoencoders (SAEs) are standard tools for mechanistic interpretability, but current SAE families are constrained by fixed encoder nonlinearities such as ReLU, JumpReLU, and…
cs.LG2026
Scalable Circuit Learning for Interpreting Large Language Models
Naiyu Yin, Dennis Wei, Tian Gao +3
A prominent research direction in mechanistic interpretability is learning sparse circuits over LLM components to reveal how they jointly produce model behavior. However, raw neuro…