Evaluation
evaluation_documented
Level: activity, page
Values
results_reported- The page states that this activity was tested, evaluated, audited or monitored and reports a result (a metric, a finding or a conclusion).
conducted_no_results- The page states that an evaluation, test, audit, pilot review or monitoring took place or is in place, but reports no result.
required_or_planned- The page states that an evaluation is required (by a rule) or planned, with no evidence that it has happened. A policy finding that requires evaluation of GenAI tools takes this value.
unclear- The page makes an evaluative statement, but its subject, or whether it concerns this GenAI activity, cannot be settled. An evaluation clearly of a different tool or activity is
not_documented, notunclear.
Missing answers
not_documented- The page has no evaluation statement about this activity.
evaluation_type_pre_deployment_testing
Level: activity, page
Values
yes- The page documents an evaluation of this type for the activity: testing before deployment.
no- The page does not document an evaluation of this type (testing before deployment) for the activity.
evaluation_type_pilot_evaluation
Level: activity, page
Values
yes- The page documents an evaluation of this type for the activity: evaluation of a pilot or trial.
no- The page does not document an evaluation of this type (evaluation of a pilot or trial) for the activity.
evaluation_type_independent_audit
Level: activity, page
Values
yes- The page documents an evaluation of this type for the activity: an audit or evaluation by an independent body.
no- The page does not document an evaluation of this type (an audit or evaluation by an independent body) for the activity.
evaluation_type_ongoing_monitoring
Level: activity, page
Values
yes- The page documents an evaluation of this type for the activity: ongoing monitoring of the tool in use.
no- The page does not document an evaluation of this type (ongoing monitoring of the tool in use) for the activity.
evaluation_type_user_feedback
Level: activity, page
Values
yes- The page documents an evaluation of this type for the activity: user feedback or satisfaction surveys.
no- The page does not document an evaluation of this type (user feedback or satisfaction surveys) for the activity.
evaluation_type_bias_or_safety_testing
Level: activity, page
Values
yes- The page documents an evaluation of this type for the activity: bias, fairness or safety testing.
no- The page does not document an evaluation of this type (bias, fairness or safety testing) for the activity.
Support
support_status_second_reason
Level: row
Values
downgraded_no_quote- The answer also failed because it has no supporting quote.
downgraded_low_confidence- The answer also failed because the classifier's certainty was below the 0.8 bar.
downgraded_quote_not_stating_value- The answer also failed because its quote does not state the value.