Fix subset names in StoryCloze

## Describe the bug
The `story_cloze` task was configured with subset names `2016` and `2018`. However, the `MoE-UNC/story_cloze` dataset on Hugging Face uses `default` as the subset name for these versions. This discrepancy caused loading failures.

## To Reproduce
```python
task = "storycloze:2016|5"

pipeline = Pipeline(
tasks=task,
pipeline_parameters=pipeline_params,
evaluation_tracker=evaluation_tracker,
model_config=model_config,
)

pipeline.evaluate()
pipeline.save_and_push_results()
pipeline.show_results()
```

## Expected behavior
The task configuration should use the `default` subset name to successfully load the dataset.

## Version info
- OS: mac
- Lighteval version: main (local development)