Tests for Feedback
Test how well AI can provide thoughtful feedback on your work and suggest meaningful improvements.
Tests in this category
-
Give feedback that contradicts content
Failure test: ask for feedback that would require contradicting factual claims. Watch for whether model catches the issue or provides confused feedback.
See 0 results → -
Review and critique a fundraising strategy
Test the model's ability to provide constructive, specific feedback on a charity's strategy. Should identify strengths and provide actionable improvement suggestions.
See 0 results → -
Critique a program design
Test ability to analyze program logic and provide feedback on design. Should identify potential issues, unintended consequences, and improvements.
See 0 results → -
Provide editorial feedback on written content
Evaluate ability to give detailed editorial feedback with specific suggestions for improvement. Should address clarity, tone, structure, and impact.
See 0 results → -
Provide feedback on ambiguous work
Failure test: provide poorly defined work with no clear evaluation criteria. Watch for whether model asks clarifying questions or gives vague feedback.
See 0 results → -
Provide feedback requiring specialized expertise
Test with highly technical content requiring deep domain knowledge. Monitor for whether model admits limitations or overreaches.
See 0 results →