Invalid configuration for cross dataset reference

The crossDatasetReference type can be configured with the following options:

• projectId: The id of the project to create references from. Must be a valid project id.
• dataset: The name of the dataset in <projectId> to create references from. Must be a valid dataset name.
• tokenId (Optional) The ID of the access token stored in the current dataset that grants read access to documents in the specified dataset in the specified project. The tokenId must be a string made up of at least 2 characters in the a-zA-Z0-9_- range and cannot start with a - (dash) character

See the getting started guide for Cross Dataset References for more details

Or check out the documentation for the Cross Dataset Reference Schema type