- Rating
- 2
- Confidence
- 4
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
This paper proposes SCOPE, a reinforcement learning framework for prompt rewriting in T2I generation.
Standard prompt rewriting RL gives a single image-level reward equally across all generated tokens, which treats split subwords as independent update targets. SCOPE asks the rewriter language model to produce semantic phrases separated by dividers during generation. It averages the policy log-probability changes within each phrase and uses them as detached weights to adjust the token-level policy loss. The authors test SCOPE with Qwen3-8B across Qwen-Image, SD3, and FLUX.1-dev on GenEval and T2I-CompBench.
Critical strengths
-
The span-level reweighting method introduces no extra visual reward models or extra rendering cost. The rewriter language model generates phrase separators together in a single pass, and the span weights reuse already computed log-probabilities without needing an external text parser.
-
The training pipeline includes a reliable fallback mechanism that protects training stability. When the model generates wrong separators, misaligned tokens, or cut-off outputs, the loss automatically falls back to standard uniform-token GRPO, with recorded fallback rates staying below 0.6% across all runs.
-
The paper is well-written and easy to follow.
Critical weaknesses
-
The performance gains over standard GRPO are marginal and show clear drops in some categories. On GenEval, the overall improvement on Qwen-Image is only 0.04, while the color score drops from 0.94 to 0.93. Across all three backbones on T2I-CompBench, the average score gains are only around 0.04, which is very small considering the randomness in diffusion generation.
-
The proposed span weighting only reflects the size of policy changes without direct credit assignment to the image reward. A phrase with a larger change between current and old policies does not necessarily contribute more to visual alignment. The method scales token weights based on probability changes without checking true visual cause, leaving the credit assignment problem unsolved.
-
The span weights are limited to a very narrow range between 0.8 and 1.2. This tight bound makes the weight change act like a mild smoothing factor on token gradients. The paper lacks tests on different ranges, so it is unclear whether the performance gain comes from real phrase weighting or simple gradient smoothing.
-
The evaluation relies entirely on automated detector scores without human verification. The paper lacks blind human studies to verify if these small score changes lead to noticeably better images.
Questions for authors
Please kindly refer to the weaknesses.
6 comments
Q3: Do semantic boundaries matter beyond general gradient smoothing, and how sensitive are the results to the clipping range?
A3: The partition ablation shows that semantic grouping matters, while the range ablation shows that moderate clipping performs best. These experiments examine different factors.
To test the partitioning scheme, Figure 3(b) compares semantic spans with token-local, random-span, and fixed-window partitions on Qwen-Image.
| Partitioning scheme | GenEval Overall |
|---|---|
| Token-local | 0.89 |
| Random spans | 0.86 |
| Fixed 3-token windows | 0.87 |
| GRPO | 0.91 |
| SCOPE semantic spans | 0.95 |
Semantic spans score 0.95, while the other partitioning schemes score between 0.86 and 0.89. Thus, grouping arbitrary adjacent tokens does not reproduce the result achieved by semantic spans.
To test the strength of reweighting, we hold the other settings fixed on FLUX.1-dev and vary the raw clipping range before normalization.
| Experiment | Raw span-weight clipping range | GenEval Overall ↑ | Single | Two | Counting | Colors | Position | Attribute |
|---|---|---|---|---|---|---|---|---|
| E1: Reweighting disabled | [1.0, 1.0] | 0.77 | 0.98 | 0.86 | 0.74 | 0.83 | 0.62 | 0.56 |
| E2: Weaker reweighting | [0.9, 1.1] | 0.81 | 0.99 | 0.86 | 0.77 | 0.86 | 0.80 | 0.57 |
| E3: Default setting | [0.8, 1.2] | 0.84 | 1.00 | 0.89 | 0.83 | 0.89 | 0.84 | 0.60 |
| E4: Stronger reweighting | [0.6, 1.4] | 0.80 | 0.99 | 0.87 | 0.75 | 0.85 | 0.79 | 0.57 |
The default range achieves an Overall score of 0.84, compared with 0.77 when reweighting is disabled, 0.81 with the narrower range, and 0.80 with the wider range. Thus, both the choice of update unit and the strength of reweighting affect performance.
Q4: Are the gains perceptible to people, and was a blind evaluation conducted?
A4: Human raters can perceive the improvement: SCOPE achieves preference rates of 74.28% over GRPO and 88.26% over the original baseline. The completed study is a paired comparison on FLUX.1-dev.
We conducted a blind evaluation: 5 raters independently assessed all 553 GenEval prompts, with 4 seeds per prompt, yielding 2,212 matched image pairs for each comparison.
Raters used the original prompt as the reference and made a forced-choice judgment with no ties. If only one image met the original prompt’s requirements, they selected that image. If both images met the requirements, they compared visual quality. If neither image met the requirements, they considered both instruction following and visual quality.
| GenEval dimension | Image pairs | SCOPE preference over GRPO | SCOPE preference over the original baseline |
|---|---|---|---|
| Single object | 320 | 60.25% | 87.75% |
| Two objects | 396 | 68.33% | 90.51% |
| Counting | 320 | 82.00% | 91.56% |
| Colors | 376 | 76.91% | 81.54% |
| Position | 400 | 87.00% | 96.60% |
| Color attribution | 400 | 70.00% | 81.80% |
| Overall | 2,212 | 74.28% | 88.26% |
Preference rates are averaged across raters, and Overall is weighted by the number of image pairs. SCOPE’s preference rates over GRPO reach 82.00% for Counting and 87.00% for Position. Thus, the gains on compositional generation are also reflected in human judgments.
Thanks for the response. A3 answers my question on the clipping range. The human study in A4 also mostly addresses my concern on human evaluation.
Two main concerns are still not resolved:
-
A2 agrees that SCOPE does not solve causal credit assignment, and says tokens in each semantic span share an update coefficient. But in uniform-token GRPO, all tokens already have weight 1.0, so each span is already consistent inside. SCOPE gives different weights to different spans, yet there is no clear reason why a span with a larger policy log-probability shift should get a larger weight. Calling it 'Credit Optimization' and 'SCOPE credit assignment' is an overclaim.
-
A1 only repeats Appendix A.13 and gives no standard deviations or multi-run variance. So it is unclear if the ~0.04 gains on T2I-CompBench are much larger than renderer noise.
So my main concerns on the method motivation and score stability remain.
Thank you for your careful review and constructive suggestions. We respond to each point below.
A1: The gains are stable under diffusion sampling because we control sampling randomness during both training and evaluation. During training, all members of each rollout group use the same image-generation seed, reducing interference from different sampling noise within a group. Different rollout groups and training steps use different seeds, so training does not rely on a single fixed seed. During evaluation, each GenEval prompt is used to generate images with 4 different seeds, and each T2I-CompBench prompt with 10 different seeds. Benchmark scores are computed from these generated images. These training and evaluation protocols are described in Appendix A.13 of the submitted paper.
A1.a: SCOPE uses policy-change signals to adjust span reweighting, so its gains depend on how much different semantic combinations help the generation task. Our inspection of rewritten prompts shows that the effective combinations learned by the model mainly concern counting layouts and the coordination of absolute positions with relative relations. These descriptions directly reinforce counting and spatial constraints, whereas color requirements depend more on correctly associating colors with objects and may benefit less from the same kinds of layout descriptions. In addition, Qwen-Image already has strong color-generation ability: after GRPO optimization, its Colors score is 0.94, leaving less room for further improvement.
A2: Policy changes use an update signal already available in GRPO training. SCOPE averages this signal within each semantic phrase, so tokens expressing the same requirement share an update coefficient. This refines the granularity of policy updates, aligning them with the semantic structure of the prompt without requiring an additional VLM or phrase-level reward model.
We agree that probability changes between the current and old policies reflect policy-update signals and do not directly measure a semantic span’s contribution to the image reward or visual alignment. SCOPE does not solve causal credit assignment at the span level. Instead, it addresses a narrower issue: an image-level advantage would otherwise be applied uniformly to every token in a rewrite. SCOPE pools policy changes within semantic spans and assigns a shared coefficient to the tokens in each span, aligning the granularity of updates with the semantic structure of the prompt.