[FEATURE] Review `setup_logging` and add control flags

[alvarobartt]

Description

Currently the logging handler created for the LLMs is named as distilabel.llm.{llm.model_name}, but the llm.model_name property shouldn't be used since it can be confusing or even expose a private URL when we're using either Inference Endpoints or an LLM subclass without a name other than the URL for example. So on, the logging handler for the LLMs should be improved.

Additionally, it would be nice to add a control flag named verbose for the steps that we want to silence or make less verbose as e.g. the data loading step which will work in most of the cases and adding logs per batch whenever there are multiple batches can overload the logs with no relevant information.

Other than that, an in detail review of the current logging approach could be done too.