Why the output of the e_vdwl is positive

Dear lammps-users:
When I used The lj/cut styles to compute the standard 12/6 Lennard-Jones potential,the output of the e_vdwl is negative.As far as I know,the distance of molecule r is always more than sigma.so the the output of the e_vdwl is negative,but I am confused why the output of the e_vdwl is negative?
Eagerly waiting for the reply, and thank you very much!

   Sincerely yours
   Sneaker Bai

Dear lammps-users:
   When I used The lj/cut styles to compute the standard 12/6 Lennard-Jones
potential,the output of the e_vdwl is negative.
As far as I know,the distance of molecule r is always more than sigma.

the lj/cut potential acts between pairs of atoms not molecules, and
those atoms can very well be closer than sigma. how much and how
frequent depends on the simulated conditions.

so the the output of the e_vdwl is
negative,but I am confused why the output of the e_vdwl is negative?

this statement makes no sense and since it is missing the entire
context it is not even possible to guess what you are asking.

axel.