Allowed topics

The Allowed topics guardrail can be applied to both user messages (Input) and the model's responses (Output).

• Allowed topics (Input): Ensures user messages remain centered around approved topics, focusing on relevant discussions.
• Allowed topics (Output): Guarantees that responses provided to users are strictly based on allowed topics, ensuring relevant and appropriate

Example:

In the following example, we've set the allowed topic to be "Food":

Configuration:

Parameters required to set up a Allowed topics guardrail:

• List of Topics: Specify the topics to focus the discussion on.
• OpenAI-Compatible Model: An OpenAI-compatible model is used to enforce the allowed topics test.

Using the UI

Using the SDK

Create Guardrail

from superwise_api.models.application.application import ApplicationAllowedTopicsGuard
from superwise_api.models.application.application import OpenAIModel

allowed_topics_guard = ApplicationAllowedTopicsGuard(
    name="Restricted topics guardrail",
    tag= "input",
    topics=["technology", "science", "education"],
    model=OpenAIModel(version="gpt-4o", api_token="REPLACE")
)
)

Add guards to application

app = sw.application.put(
    str(app.id),
    additional_config=AdvancedAgentConfig(tools=[]),
    llm_model=model, 
    prompt=None, 
    name="App name",
    show_cites = True,
    guards=[ApplicationAllowedTopicsGuard]
)