Reviewing and Sharing Results

Use Task Results to find one measurement, inspect its evidence, record context, and involve the project when follow-up is needed.

Find a Result

1. Open Task Results.
2. Filter by chip, execution, task, target, or status.
3. Select a row to preview its figure, message, and parameters.
4. Open the full result when you need artifacts, history, notes, discussion, or re-execution.

Start from Execution instead when task dependencies or run-wide progress matter. Start from Chip when you want the latest result for one task across the topology.

Inspect the Evidence

Review the status and message first, then compare the input, run, and output parameters. Open the figure and download raw data when the task produced it. Check the execution and target before comparing this result with another run.

Use Calibration Data and Provenance when you need parameter history or the lineage of a saved value.

Record Result Context

Add a result note for context that belongs to this measurement, such as an anomaly or the reason for a parameter choice. Use a pinned dashboard summary when the context should remain visible for a chip, qubit, or coupling in a selected cool-down or time range.

Notes are shared with the project. Write what another member needs to interpret the result, not a private reminder that lacks the chip, target, or decision context.

Start a Discussion

Select New thread from the result when a finding needs a decision or follow-up. QDash fills in the result, chip, and target so the discussion links back to its evidence.

Use Forum to reply, mention another member, update the thread status, and close the thread when the work is resolved. Create a standalone forum thread for topics that span multiple results or concern project-wide policy.

Reuse the Finding

Move a conclusion into the task's knowledge when it should guide future analysis of the same task. Knowledge is read-only: changes go through the task-knowledge repository, so one person's edit does not change what everyone else's evaluations and chats read.

1. On Knowledge, open the task and select Download zip.
2. Update the task background, expected plot, common failures, parameters to extract, or examples in the unzipped task.md and examples/. Keep the task title and required section headings unchanged, and preserve the YAML frontmatter identity fields.
3. Zip the folder again and evaluate with it as a knowledge draft; see Try a Knowledge Change Before Proposing It.
4. Propose the change as a pull request to the task-knowledge repository. After it merges, the operator runs task knowledge-pull to update the deployment.

Request an AI Review

Select Ask AI on one result to start a recorded evaluation or open a chat with that result's context. Select Evaluate on Chip to evaluate the displayed task across eligible targets.

AI output is supporting evidence, not an automatic calibration decision. Confirm it against the figure, raw data, task knowledge, and current calibration context. See AI Evaluations for the complete workflow.

Follow Notifications

Mentions, replies, and system events arrive at the bell in the top bar, which shows the unread count. Select the bell to preview the five newest notifications and open one, or select View all to open Inbox with the full list. Opening a notification marks it read; Mark all read clears the rest. Open the linked record to recover its project and result context.

Share and Hand Off a Result

Before sending a QDash link, confirm that the recipient belongs to the active project. Project membership controls access; removing a member revokes access without deleting the records they produced.

When another member will continue the calibration:

1. Confirm the active project and chip.
2. Open Task Results and share the result link or start a forum thread from the result.
3. Record the execution, target, evidence, and next decision in the result note or thread.
4. Ask the recipient to confirm the same project and chip before starting the next task or workflow.

Saved calibrated outputs become the current state for later tasks in the same project. An exploratory run with Save calibrated outputs to DB turned off remains available for review, but later tasks do not use its outputs as current values. See Task Parameters for that choice.