## Dataset Structure

Each record has four top-level fields:

- **input**: Data to recreate the example in your application (required).
- **expected**: Ideal output or ground truth (optional but recommended for evaluation).
- **metadata**: Key-value pairs for filtering and grouping (optional).
- **tags**: Labels for organizing and filtering records (optional).

## Where to go from here

- [Create datasets](/content/docs/annotate/datasets/create/index.html) from uploads, the SDK, production logs, user feedback, traces, or Loop.
- [Build dataset pipelines](/content/docs/annotate/datasets/pipelines/index.html) to transform project logs into dataset rows in bulk.
- [Manage datasets](/content/docs/annotate/datasets/manage/index.html) — tag and star, save snapshots, define schemas, customize table views, and edit records.
- [Use in evaluations](/content/docs/annotate/datasets/use-in-evaluations/index.html) by passing datasets to `Eval()`, assigning them to environments, or converting experiment results.
- [Track performance](/content/docs/annotate/datasets/track-performance/index.html) to see which experiments used a dataset and how each row performs.

For human review workflows on dataset records, see [Human review](/content/docs/annotate/human-review/index.html) and [Custom views](/content/docs/annotate/custom-views/index.html).