Add a note about data-aware mixed precision assignment

[nikita-savelyevv]

What does this PR do?

Recently there appeared a confusion regarding when the dataset argument actually affects the resulting model. I've added some notes to help clarifying this.

Namely, if users provide compression ratio<1 and a dataset at the same time, according to NNCF data-aware mixed precision using MAX_ACTIVATION_VARIANCE criterion will be applied. https://github.com/openvinotoolkit/nncf/blob/develop/nncf/quantization/algorithms/weight_compression/algorithm.py#L96

Otherwise, if no data-aware compression algorithm is selected, dataset argument is not used.

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?

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[helena-intel]

Thanks @nikita-savelyevv !

[echarlaix]

Thanks for the addition @nikita-savelyevv !