> ## Documentation Index
> Fetch the complete documentation index at: https://braintrust.dev/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Human review output format: string vs JSON

export const plans_0 = "Any"

export const deployments_0 = "Any"

export const data_plane_version_0 = undefined

export const use_case_0 = "Use case - Understanding how human review values are stored in the expected field"

<Note>
  **Applies to:**

  * Plan - {plans_0}
  * Deployment - {deployments_0}
  * {data_plane_version_0}
  * {use_case_0}
</Note>

## Summary

**Behavior:** When labeling samples with human review, the value written to the `expected` field format depends on the number of reviews applied to the row.

## Expected field formats

### Single human review

When only one human review is configured, the output is a plain string:

```json theme={"theme":{"light":"github-light","dark":"github-dark-dimmed"}}
"pass"
```

### Multiple human reviews

When two or more human reviews are configured, all values are stored as JSON:

```json theme={"theme":{"light":"github-light","dark":"github-dark-dimmed"}}
{
  "review name": "pass",
  "second review": "value"
}
```

## Working with different formats

### Force JSON format

To ensure consistent JSON output, add a second human review to your project and apply both reviews to each row.

### Manual expected field updates

You can manually edit the `expected` field in the UI to convert between string and JSON formats as needed.

## Notes

* This behavior is consistent across human review configurations
* The format is determined when the review is submitted and cannot be changed through project settings
* See the [human review guide](/annotate/human-review) for general configuration reference