LLMFilter#
classDiagram
FilterBase <|-- LLMFilter
Monitorable <|-- FilterBase
- class council.filters.LLMFilter(llm: LLMBase, filter_on: List[str] | None = None)[source]#
Bases:
FilterBase
Filter using an LLM to filter chain responses.
- __init__(llm: LLMBase, filter_on: List[str] | None = None) None [source]#
Build a new LLMFilter.
- Parameters:
llm – model to use for the filtering.
filter_on – List of filters to filter chain responses on.
- execute(context: AgentContext) List[ScoredChatMessage] #
Selects responses from the agent’s context.
- Parameters:
context (AgentContext) – The context for selecting responses.
- Returns:
A list of scored agent messages representing the selected responses.
- Return type:
List[ScoredChatMessage]
- Raises:
None –
- render_as_dict(include_children: bool = True) Dict[str, Any] #
returns the graph of operation as a dictionary
- render_as_json() str #
returns the graph of operation as a JSON string