README.md

---
license: cc-by-sa-4.0
---

# Dataset Card for RedirectQA v0.1.0
RedirectQA v0.1.0 is an entity-based Question Answering (QA) dataset that considers multiple surface forms.
For each triplet, the dataset includes multiple QA pairs based on different surface forms, allowing for a more comprehensive evaluation of entity knowledge of models.

## Dataset Structure
```Python
DatasetDict({
    train: Dataset({
        features: ['subject', 'predicate', 'object', 'subject_surface', 'predicate_surface', 'object_surface', 'subject_aliases', 'object_aliases', 'subject_wiki_title', 'object_wiki_title', 'question', 'possible_answers', 'triplet_id', 'subject_redirect_wiki_title', 'subject_surface_category', 'index'],
        num_rows: 22869
    })
})
```

## Example
```Python
from datasets import load_dataset

dataset = load_dataset("naist-nlp/RedirectQA", revision="v0.1.0")
```