Validator: Improve Topic Relevance support

Describe the current behavior
The frontend currently supports the Topic Relevance (LLM) validator, but there are a couple of backend limitations:

1. Configuration dropdown support

• The existing API used to fetch configurations for the Topic Relevance validator does not support the Topic Relevance (LLM) validator.
• When attempting to fetch configurations for Topic Relevance (LLM), the API returns a “validator not supported” error.
• As a result, users cannot select from existing configurations through a dropdown, unlike the standard Topic Relevance validator.

2. Create new configuration support

• The workflow for creating a new configuration is currently supported for the Topic Relevance validator only.
• Topic Relevance (LLM) configurations cannot be created, managed, or selected through the existing backend APIs.

Describe the enhancement you'd like
Add backend support for Topic Relevance (LLM) configuration management, including:

1. Support fetching available configurations for the Topic Relevance (LLM) validator through the configuration API.
2. Support creating new Topic Relevance (LLM) configurations.
3. Support listing, retrieving, and managing Topic Relevance (LLM) configurations in the same way as the existing Topic Relevance validator.
4. Ensure the APIs return data in a format compatible with the existing frontend configuration dropdown and creation workflows.

Additional context
The frontend implementation for the Guardrails UI updates assumes that Topic Relevance (LLM) behaves similarly to the existing Topic Relevance validator.

Current blockers observed:

• Fetching configurations for Topic Relevance (LLM) returns a “validator not supported” error.
• Creating new configurations for Topic Relevance (LLM) is not currently supported by the backend APIs.

The goal is to bring Topic Relevance (LLM) to feature parity with the existing Topic Relevance validator from a configuration management perspective.

Start by tracing the existing configuration API and creation workflow for the standard Topic Relevance validator, then locate where Topic Relevance (LLM) is rejected as unsupported. Compare the two validator paths and verify that fetching, creating, listing, retrieving, and managing LLM configurations return data compatible with the frontend workflows.