Answer from cs61c-cm (Christopher Marcellino 15775059) for Question 2
Using the low bits except the byte addressing bits increases the hit rate due to spatial locality. The low two bits are not used since MIPS cache word size is 8 bytes

Due to the spatial locality of data and instructions, using high-order bits would cause a very high number of cache misses, since only one word per a large section of memory (addresses that start with those bits) can be cached.