IDEEA: training-free Input-Dependent stEEring via Activation cluster matching
IDEEA (Input-Dependent stEEring via Activation cluster matching via Activation cluster matching), a training-free framework for input-dependent steering that aligns the model toward the target concept while preserving the input's original representation.
Zheng Wang, Mu-Chen Li, Ren-Jie Liao et al.
· 0 citations