Circuit Condensation can reduce mechanistic interpretability circuits by 8.1x on average and up to 316x compared to frozen baselines across four behaviors and eight models

Across four behaviors and eight models, condensed circuits are smaller than the strongest frozen baseline in 30 of 32 settings, by $8.1\times$ on average and up to $316\times$.

http://arxiv.org/abs/2608.27254v1