# NER Benchmark

Named Entity Recognition benchmark dataset.

## Entities

- PER (Person)
- ORG (Organization)
- LOC (Location)
- MISC (Miscellaneous)

## Task

Predict entity labels for each token.

## Metrics

- Precision
- Recall
- F1 (entity-level)

## Format

```json
{
  "tokens": ["Apple", "is", "great"],
  "ner_tags": ["ORG", "O", "O"]
}