How strongly the model adapts to the training labels. Regularization strength. Lower value means the model fits more heavily to the training data