I can't swear exactly how the Zero BMS's behave, but I did build an electric bicycle a while back and I read up pretty thoroughly on the operation of the BMS I used for that battery pack.
For bulk charging, a high-power switch (a MOSFET in this case) connects the charger directly to the battery pack, and all the cells are charged in series (that pack had only a series string, no parallel hookup). This gives all the cells exactly the same charging current, but since no two cells are ever truly identical, one of them will hit 100% capacity first, and when that happens, the high-current charging path is opened up. Battery charger voltage is then applied to a series resistor string, one for each cell, but the current through each resistor can be diverted by a switch (smaller MOSFETs, two per cell) through each individual cell instead of through the resistor. Each cell "decides" whether it needs to continue charging or if it's done. The current through the resistor string is far smaller than the main charging current (to keep heat generation down), and will charge each cell only very slowly. Eventually all cells will achieve the same voltage, the maximum safe voltage for that cell, and the whole pack is considered fully charged.
If this is similar to Zero's BMS operation, which I suspect it is, there is no equalization done until the bulk charging phase is turned off. I think Kocho's right, during the bulk charging phase, the cell voltages may equalize some but it's just due to the cells all approaching "fully charged" voltage together. Cell voltage is pretty stable until the end of the discharge cycle, where the difference between a tired cell and a not-so-tired cell increases much faster.