Prompt: This is a cars attributes dataset. I have a hidden car in mind. You are allowed to ask me binary questions about the hidden car's attributes, for you to figure out the hidden car. 
After receiving an answer, at each step, you must print what your current belief distribution is and the calculated entropy drop, about the possible hidden car. 
We can stop when there's no more distinguishing attributes between the leftover options. You're not allowed to solve this using programming.
