[OpenVINO] Adopt new mxfp4 quantization logic

[nikita-savelyevv]

What does this PR do?

Changes:

• Now with NNCF develop version there is an explicit "mxfp4" compression mode. The group size value is fixed to 32 for it according to mxfp4 definition.
• Remove temporary cb4 pre-release logic introduced in Add support for cb4_f8e4m3 quantization mode. #1378.

There are also other compression modes "mxfp8_e4m3", "fp4_e2m1" and "nvfp4" added to NNCF, but they are planned to be added to optimum-intel after the next NNCF release.

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you make sure to update the documentation with your changes?
• Did you write any new necessary tests?

[nikita-savelyevv]

Delegate group size value selection to NNCF. This is backward compatible in terms of the default value since NNCF also selects -1 for 8-bit types and 128 for 4-bit non-mxfp4 types by default.

[nikita-savelyevv]

cc @ljaljushkin @daniil-lyakhov

[echarlaix]

LGTM, thanks @nikita-savelyevv !