Hi, Mayank,
Sorry for the late response: the world got turned on its head no so long ago.
As mentioned in Sec. 7.1 of the manual, the input files used to generate the output you see can be found in the INDUS repo under indus/examples/bulk_water. I cannot speak to exactly why you see what you do, but it strikes me that your settings are not actually applying a bias on Ntilde-v, since its value is still roughly the same as at equilibrium in the unbiased ensemble.
What extra information would it be helpful to add to the manual/example? In the bulk water example, I believe I have added all of the information needed to reproduce the given results using any MD engine supported by PLUMED. While I do not provide my GROMACS input files, such information would be of little or no use to someone using e.g. LAMMPS or AMBER.
Sean