frontier models
The Image Editor Became Its Own Strictest Critic
A frozen internal critic rejected edits that looked plausible but missed the request or damaged preserved content.
Summary
A frozen internal critic rejected edits that looked plausible but missed the request or damaged preserved content.
Rubric-CEPR improves an image editor using its own generations rather than human-edited targets or an external reward model. A planner proposes edits, the editor creates candidates, and a frozen critic applies separate checks for making the requested change, removing the old state and preserving everything else. Non-compensatory gates reject candidates that fail any essential condition before adapter training distills the best survivors. On Qwen-Image-Edit, the authors report an ImgEdit increase from 4.36 to 4.60 and a 24.9% gain on object isolation; a second editor improved 7.8%. These are benchmark gains from self-distillation, not a guarantee against editing errors.
Why it matters
A frozen internal critic rejected edits that looked plausible but missed the request or damaged preserved content.
Limits and context
- These are benchmark gains from self-distillation, not a guarantee against editing errors.
Key claims
A frozen internal critic rejected edits that looked plausible but missed the request or damaged preserved content.
Qualification: These are benchmark gains from self-distillation, not a guarantee against editing errors.
Evidence: source-2026-10-10-004
Sources
- arXiv preprint 2610.12469, original v1arXiv / submitting research authors · primary research
Corrections
No corrections have been recorded for this story.