Dashboard

PXChat Evals

Overview Cases Run Builder Presets RunsHelp

New case

PX Chat / Golden Master

Create a new eval case directly in the harness.

Case keySource case IDCase typeWorkflow statusSource statusQ-ID (optional)Scenario keyScenario turnScenario name (JA)Scenario name (EN)

Tags

Type to filter existing tags. Press Enter to add the highlighted tag.

Release criticalArchived

Check exact emitted signals for this case

Leave unchecked to skip signal grading. Check it with no selection to expect no signal.

Prompt (JA)Expected (JA)Prompt (EN)Expected (EN)Target (JA)Target (EN)