All of the above CPUs run at the same Voltage of 1.6V. The big advantage of 2.4W was achieved by using on-die level two cache in comparison to discrete level two cache chips.

Those Pentium II processors with on-die level two cache, using the 'Dixon'-core, are officially called Pentium II PE. PE stands for performance enhanced. The discrete 512kB level two cache of the Pentium II processor is clocked with half the speed of the CPU core, while the 256kB on-die cache of the Pentium II PE processor runs at CPU clock.

The Celeron at MMC-1 started right away with 128kB on-die level two cache.

This is no error. The Celeron processor was available at higher clock rates then the Pentium II processor. You can clearly see the reason for this in the above chart. The good old overclocker tip: increase the voltage and it will run faster. However, the processor also got much hotter too.