EvalResult Type
The result of evaluating a single case
Record fields
| Record Field |
Description
|
Full Usage:
ActualOutput
Field type: string
|
The agent's actual output
|
Full Usage:
CaseId
Field type: string
|
The eval case that was run
|
Full Usage:
EvaluatorName
Field type: string
|
Evaluator that produced this result
|
|
Optional execution trace for deeper analysis and regression detection
|
Full Usage:
LatencyMs
Field type: int64
|
Time taken to get the agent's response (ms)
|
Full Usage:
Reason
Field type: string
|
Reason/explanation for the verdict
|
|
Timestamp of evaluation
|
|
The evaluation verdict
|
Nao