DataDiscrete.get_activation_outhead DataDiscrete.get_activation_outhead() → str | torch.nn.Module | None[source] Returns the activation function for approximating this feature.