Skip to main content

ConceptExplanation

Node-level explanation of a concept execution result.

Populated when ExecuteRequest.explain=True. Null on responses where explain was False or omitted.

output — final computed value; matches ConceptResult.value. contributions — per-signal attribution: signal_name → contribution_amount. Values are dimensionless; their sum equals 1.0 for normalised concepts, or is unbounded for raw aggregations. nodes — per-node computation records in topological execution order. trace — step-by-step trace with intermediate values (debug mode only).

output objectrequired
anyOf
number
contributions object
property name*number
nodes object[]
  • Array [
    • node_idNode Id (string)required
    opOp (string)required
    inputsobjectrequired
    paramsobjectrequired
    output_value objectrequired
    anyOf
    number
    output_typeOutput Type (string)required
  • ]
    • traceobject[]
    ConceptExplanation
    {
      "output": 0,
      "contributions": {},
      "nodes": [
        {
          "node_id": "string",
          "op": "string",
          "inputs": {},
          "params": {},
          "output_value": 0,
          "output_type": "string"
        }
      ],
      "trace": [
        {}
      ]
    }