Copy as Markdown[Open in ChatGPT](https://chatgpt.com/?q=Read%20https%3A%2F%2Fcoralogix.com%2Fdocs%2Fuser-guides%2Fdata-layer%2Fdefault-dataspace%2Fai-evaluations.md%20and%20help%20me%20with%20my%20question%20about%20this%20Coralogix%20documentation%20page.)[Open in Claude](https://claude.ai/new?q=Read%20https%3A%2F%2Fcoralogix.com%2Fdocs%2Fuser-guides%2Fdata-layer%2Fdefault-dataspace%2Fai-evaluations.md%20and%20help%20me%20with%20my%20question%20about%20this%20Coralogix%20documentation%20page.)

# `ai.evaluations`

## Purpose[​](#purpose "Direct link to Purpose")

The `ai.evaluations` dataset is the GenAI evaluation dataset. It stores evaluation results for your AI applications, including the [user feedback scores](https://coralogix.com/docs/user-guides/ai/explorer.md#send-user-feedback-scores) you send. Coralogix ingests each feedback log into it, so you can query scores directly in DataPrime and build custom views, dashboards, or alerts on top of them.

`ai.evaluations` is a [predefined dataset](https://coralogix.com/docs/user-guides/data-layer/default-dataspace/predefined-datasets.md): Coralogix provisions it in the `default` dataspace, and your AI telemetry fills it.

## Query the dataset[​](#query-the-dataset "Direct link to Query the dataset")

```
source ai.evaluations

| filter logRecord.attributes['gen_ai.evaluation.name'] == 'user_feedback'
```

## Schema description[​](#schema-description "Direct link to Schema description")

Every record carries the standard log record envelope: engine metadata in `$m`, labels in `$l`, and the evaluation payload in `$d`.

| Field                 | Type               | Required | Description                                                                                              |
| --------------------- | ------------------ | -------- | -------------------------------------------------------------------------------------------------------- |
| `$d`                  | object             | Yes      | Entity-type-specific payload — shape defined by each entity type's schema                                |
| `$l`                  | map                | Yes      | User labels: free-form string-to-string map                                                              |
| `$l.<key>`            | string             | No       |                                                                                                          |
| `$m`                  | object             | Yes      | Engine metadata                                                                                          |
| `$m.cxEventId`        | string (uuid)      | Yes      | Globally unique event id                                                                                 |
| `$m.ingressTimestamp` | string (date-time) | Yes      | Time the event was ingested                                                                              |
| `$m.priorityClass`    | string             | Yes      | Defines the feature-set enabled for the event<!-- --> <!-- -->One of: `high`, `medium`, `low`.           |
| `$m.severity`         | string             | Yes      | Event severity level<!-- --> <!-- -->One of: `Verbose`, `Debug`, `Info`, `Warning`, `Error`, `Critical`. |
| `$m.timestamp`        | string (date-time) | Yes      | Event timestamp                                                                                          |

## Next steps[​](#next-steps "Direct link to Next steps")

* See where feedback scores appear and how to investigate them in [AI Explorer](https://coralogix.com/docs/user-guides/ai/explorer.md#query-feedback-scores).
* See the other datasets Coralogix provisions in [Predefined datasets](https://coralogix.com/docs/user-guides/data-layer/default-dataspace/predefined-datasets.md).