[Dataset Request] Index ImmigrationQA for U.S. Immigration Law QA benchmarking (arXiv:2605.30589)

Hi there,

I'm the author of a paper I recently uploaded to arXiv and I'd love to get the dataset properly indexed on Papers with Code so the benchmark results can be tracked.

### Paper
- **Title:** ImmigrationQA: A Source-Grounded Dataset and Small-Model Adaptation for U.S. Immigration Law
- **arXiv:** https://arxiv.org/abs/2605.30589
- **Author:** Nazarii Shportun

### About the Dataset
ImmigrationQA is a source-grounded QA dataset built on top of U.S. immigration law — covering official policy documents, federal regulations, and procedural guidance. It's designed to support retrieval-augmented and fine-tuning approaches for legal QA in a high-stakes domain.

- **Size:** 17,058 QA pairs across 13 immigration subdomains
- Each answer is grounded in a specific source passage from the official documents

### Evaluation
We evaluated models using an LLM-as-judge approach on a held-out 101-example split, scoring answers from 0–3. Here are the results reported in the paper:

| Model | Mean Score (0–3) |
|---|---|
| Fine-tuned Llama 3.2 3B Instruct (LoRA) | ~1.08 |
| Llama 3 8B base (zero-shot) | ~0.85 |
| Claude Sonnet (zero-shot) | ~1.52 |

### What I'm Asking
Could you index **ImmigrationQA** as a dataset in Papers with Code and link it to the paper above under the **Question Answering** task? That would make it possible to properly register the benchmark results and allow comparison with other QA work.

Happy to provide any additional info if needed. Thanks so much!