How to keep equilibration of an unstable and easily crystallizing liquid?

Dear lammps users,
I’m simulating a liquid-liquid phase transition of Si, and I want the system to be fully equilibrated at a given point during the phase transition. However, this is not easy for me, since the system is quite unstable, and easily crystallizes during and after the liquid-liquid phase transition. Do you have any good ideas to avoid the crystallization and fluctuation, and prolong the equilibration time for my liquid-liquid phase transition ?
Bests,
Bo

Bo,

This is not a LAMMPS-related question but an
application/problem-related question.

I am sure others will have more constructive suggestions, so my only
advices are to figure out why the liquid phase is so unstable under
the conditions you imposed, whether this instability is expected by
the empirical potential used, and under what conditions should or
should not recrystallization occur.

Cheers,
Ray