Policy Debugging: Selecting a row in the Results table will now highlight assertions which evaluated to true for that row’s content.
Note: Policy lines that are inside an UNLESS or EXCEPT WHEN block will not be highlighted. Support for these lines will be added in our next release.
Test Metrics History: Users are now able to compare the metrics from the current and previous test runs.
Improvements:
Datasets: Fixed an issue with bulk editing that prevented users from bulk removing a label from rows in the content library or a dataset. When bulk editing a number of rows, the bulk editing interface will now show all labels that are common to all the rows selected, allowing you to remove those labels from all selected rows.
Test results table: We’ve updated the Matched column heading to “Label match” to make it clearer what the TRUE/FALSE values in this column mean. We’ve also added a tooltip to explain the logic, which will appear when you hover over a TRUE/FALSE badge.
For reference, the logic for the TRUE result is now:
When user labels are specified, a label match has only occurred when those user labels are a subset of “policy labels”.
When user labels are empty (i.e., content should not match any label), a label match has only occurred when the “policy labels” are also empty.
Did this answer your question?Thanks for the feedbackThere was a problem submitting your feedback. Please try again later.
