This publication covers adult-oriented AI systems. Public pages contain no explicit imagery or generated pornography.

Uncensored Index

Corrections

How factual, scoring, and attribution errors are fixed and logged.

What qualifies

Wrong model IDs, prices, dates, policy claims, test classifications, source attributions, or scoring calculations are corrections. A vendor disagreement with a documented result is not automatically an error.

Process

  1. Preserve the reported evidence.
  2. Reproduce the issue.
  3. Correct the source data or code.
  4. Rerun affected tests and pages.
  5. Add a visible correction note when the conclusion changed.