[BUG] Spec for ManagedIdentityToken

### API Spec link

https://github.com/Azure/azure-rest-api-specs/blob/166c3a2ace3d2511cbdeeada95826cd67a45895e/specification/solutions/resource-manager/Microsoft.Solutions/preview/2023-12-01-preview/managedapplications.json#L2849-L2881

### API Spec version

2023-12-01-preview (also present in 2018-09-01-preview...so since forever)

### Describe the bug

Calling listTokens, only the `authorizationAudience` and `resourceId` fields are in camel case, the others fields are all implemented as snake case. The spec has this all fields formatted as camel case meaning downstream users of these specs cannot decode the token.

Documentation matching reality: https://learn.microsoft.com/en-us/azure/azure-resource-manager/managed-applications/publish-managed-identity#accessing-the-managed-identity-token

### Expected behavior

Downstream library users can decode the response correctly, https://github.com/Azure/azure-sdk-for-python/pull/40268

### Actual behavior

`azure.mgmt.managedapplications.models.ManagedIdentityToken` returns most of its attributes as `None`

### Reproduction Steps

By doing a listTokens, you will receive:
```{"value":[{"authorizationAudience":"https://management.azure.com/","resourceId":"/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.Solutions/applications/{applicationName}","access_token":"...","expires_in":"86399","expires_on":"1743242070","not_before":"1743155371","token_type":"Bearer"}]}```

Note the formatting of the fields does not match the spec.

### Environment

Python 3.11 and https://learn.microsoft.com/en-us/python/api/azure-mgmt-managedapplications/azure.mgmt.managedapplications?view=azure-python-preview (version 1.0.0b1, the latest)

Inspect specification/solutions/resource-manager/Microsoft.Solutions/preview/2023-12-01-preview/managedapplications.json at lines 2849-2881, starting with the ManagedIdentityToken schema and the listTokens response. Compare its property names with the documented response and the reproduced payload; done means the schema represents the returned token fields so downstream SDK users can decode them correctly.

Written by the indexing model from the issue text.