DataInt.get_activation_outhead

DataInt.get_activation_outhead() str | torch.nn.Module | None[source]

Returns the loss function for approximating this feature.