During a tree's induction using the ID3 algorithm, assume the root note of the tree is at level 0. This root node represents a dataset (D) of size 10. The root node's direct child nodes are at level 1 and they represent datasets (D1 and D2) of sizes 5 and 5, respectively. D3 is a dataset of size 3 that is represented by a direct child of the node that represents D2. In D3, 1 and 2 instances are of target class A and B, respectively. The ID3 algorithm is provided with D3 and d (the current set of descriptive features) as input. If the maximum permissible depth of the tree has not been reached and the ID3 algorithm returns a leaf node with a class label of B, how many descriptive features are used to represent the instances in D?