Skip to main content

Interaction Map

Action Reference

ActionPurposeTypical Outcome
Start New ChatCreate a fresh conversation paneClean context for tests
Select ModelBind pane to model/endpointTargeted inference behavior
Send PromptTrigger model responseGenerated output + metrics
Edit & ResendRefine existing promptBetter output with same context
Retry ResponseRe-run generation quicklyAlternative response candidate
Copy MessageReuse prompt/output externallyFaster collaboration
Rate ResponseCapture quick quality signalLightweight feedback trail
Open HistoryRevisit prior sessionsContinuity and reproducibility

Parameter Controls

Common controls available in settings:
  • Temperature
  • Response length/token limits
  • Repeat penalty
  • Stop strings
  • Structured output preferences

Notes

  • Keep a naming convention for saved sessions.
  • Use a stable prompt template for A/B comparison runs.