Can be loaded via e.g.:

```python
from datasets import load_dataset
d = load_dataset("Muennighoff/xP3x-sample", "apps")
```


1,000 rows from random languages and splits of xP3x for each of the multilingual datasets represented in [xP3x](https://huggingface.co/datasets/Muennighoff/xP3x).