Interface EvaluationItem

Represents a single turn in an evaluation dataset converted from a session.

interface EvaluationItem {
    expected_intermediate_agent_responses: IntermediateAgentResponseItem[];
    expected_tool_use: ToolUseItem[];
    query: string;
    reference: string;
}

Properties

expected_intermediate_agent_responses

expected_intermediate_agent_responses: IntermediateAgentResponseItem[]

expected_tool_use

expected_tool_use: ToolUseItem[]

query

query: string

reference

reference: string