Usually in IEEE representation we consider only normalised form but if told in the question to consider non normalised form then only we shouldn't normalise it .
Here in this question to represent into 7-bit floating-point representation we use Normalized form.
The reason why we should consider only normalised form is we only represent +-1.M*2 , we don't represent number before precision point ....so it assumes 1 is represented by default.
In this question you need to represent that 0.375 base 10 to that 7-bit floating-point representation only .