[BENCHMARK] Add GenEval benchmarks

[davidberenstein1957]

Description

Add GenEval benchmark with 6 subcategories for compositional evaluation.

Details

• Source: GitHub JSON from djghosh13/geneval repo
• Subcategories: single_object, two_object, counting, colors, position, color_attr
• Collate: prompt_with_auxiliaries_collate

Implementation

• Add setup_geneval_dataset in src/pruna/data/datasets/prompt.py
• Support category param for filtering subcategories
• Register in base_datasets
• Add BenchmarkInfo entry with metrics: ["qa_accuracy"], subsets list
• Auxiliaries should include questions list and tag for evaluation
• Add test

Acceptance Criteria

• PrunaDataModule.from_string("GenEval") works (all subcategories)
• PrunaDataModule.from_string("GenEval", category="counting") works
• Auxiliaries include questions and tag fields
• Test passes