Last reviewed:
Three cheap cross-checks
- Re-run the same prompt in a fresh session and compare
- Re-run with the opposite framing ('argue against this')
- Send the output to a human who would lose if it were wrong
Related
Keep exploring
All AI planning methods, frameworks, workshops, prompt patterns, and collaboration models in one place.
Back to AI Planning hub →