LiveCallJudgment

One judgment of a Call that is not an evaluation Call (1.63.0): the suite's rubric judged over its stored transcript under live_call_judging, advisory and never a Call control. Content only in each item's rationale, under include.

An object with 19 fields.

Fields

Field	Type	Required	Description
`advisory`	boolean	yes	True while the suite's judge is not calibrated (`promptEvalJudgeAdvisory`); a judgment never controls a Call either way.
`arm`	ExperimentArmKey or null	yes	The arm the experiment assigned.
`call_id`	CallId	yes
`completed_at`	Timestamp or null	yes
`created_at`	Timestamp	yes
`environment`	Environment	yes
`error_code`	string or null	yes
`experiment_id`	ExperimentId or null	yes	The experiment whose suite judged the Call; null for no experiment.
`id`	LiveCallJudgmentId	yes
`job_id`	JobId or null	yes	The `call.judge` job that runs it.
`judge`	StageRef or null	yes	The judge that served it; null before.
`object`	`"live_call_judgment"`	yes
`outcome`	PromptEvalOutcome or null	yes	`passed` when every `must` and `must_not` item is judged `pass` or `not_applicable`; null until settled.
`rubric`	array of PromptEvalRubricVerdict	yes
`score`	number or null	yes	The share of the applicable items judged `pass`; null until `completed`.
`scrubbed_at`	Timestamp or null	yes	When the retention scrub nulled the rationales with the Call's transcript.
`status`	LiveCallJudgmentStatus	yes
`suite_id`	PromptEvalSuiteId	yes
`suite_revision`	integer	yes