: These features are typically extracted from deep layers of a neural network (such as the last fully connected layer of a pretrained VGGNet or similar architecture) to capture complex abstract information.
: ACE introduces learnable gating mechanisms in the model's cross-attention layers, which are fine-tuned per concept using these deep feature representations. ace.AT_Blacked.1.var
: The variable represents a specific semantic direction that the ACE method attempts to remove or "erase" to prevent the model from generating undesirable images. : These features are typically extracted from deep