EvalCase Type
A single evaluation test case
Record fields
| Record Field |
Description
|
Full Usage:
Description
Field type: string
|
Human-readable description
|
Full Usage:
Expected
Field type: string option
|
Expected output or reference answer (used by some evaluators)
|
Full Usage:
Id
Field type: string
|
Unique identifier for this case
|
Full Usage:
Input
Field type: string
|
The input to send to the agent
|
Additional metadata for evaluators
|
|
Full Usage:
Tags
Field type: string list
|
Tags for categorization and filtering
|
Nao