feat: RAG QA Dataset v2

[jalling97]

User Story

As an evaluator of LFAI
I want to have metrics using in depth Question/Answer evaluations
So that I can measure RAG improvement

Additional context

The LFAI RAG qa v1 dataset is a useful dataset for measuring simple question and answer pairs, but a harder task is needed in order to provide better evaluations for LFAI.

As such, a new dataset is needed. This dataset will be different in the following ways:

• The v1 version has only 8 documents in the dataset. v2 Will require more, at least double.
• The type of documents had an initially varied scope to determine what topics may be easier to answer questions surrounding. The v2 version will be more narrowed in topic.