![[Experiment] Flat Chest Prompts (Part 3) | Maturity Reinforcement, Emphasis, and Additional Testing](/tips/flat-chest-descriptor-test-part3/a00_control_front_s1_0_0000_2053695854357871006.webp)
Conclusions
- All 4 maturity reinforcement conditions failed — since
flat chestitself doesn’t work, the combined testing withmature womanetc. was moot from the start - All 4 emphasis test conditions failed — weighting (
(flat chest:1.5)), stacking, repetition (flat flat flat chest), and negative expressions all had no effect - Weighting syntax causes text artifacts —
(flat chest:1.5)caused string characters to appear in the background 1boyhas destructive side effects — the female chest stays the same but a separate male figure is added to the frame- Negative expression
not big breastsrisks backfiring —big breastsbeing recognized as a token may push toward larger breasts
This article is Part 3 of the hub article on flat chest / small chest prompt testing. It covers 14 conditions (78 images) of maturity reinforcement (Group F), emphasis testing (Group G), and additional testing (Group H).
Control (Reference)
Control (no variable) — front view:
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
Control (no variable) — side view:
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
Default output without breast size specification. Medium to slightly large size (C–D cup equivalent), spherical to teardrop shape.
large breasts — front view:
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
large breasts stably generates clearly larger breasts in 3/3. Confirmed that the “make bigger” direction works.
Maturity Reinforcement (Group F — All 4 Failed)
The strategy of “use maturity keywords to cancel any infantilization risk from flat chest” was tested. Result: since flat chest itself doesn’t work, this testing was moot from the start.
mature woman, flat chest (front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
adult woman with naturally flat chest (front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
office worker, flat chest, business casual (front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
All show the same breast size as control. mature woman has a slight aging effect on the face but is unrelated to breast size. office worker, business casual conflicts with topless and produces side effects like jacket-wearing figures. Note: since 32yo japanese actress already prevents infantilization, maturity reinforcement keywords are largely unnecessary.
Emphasis Testing (Group G — All 4 Failed)
(flat chest:1.5) — No effect + text artifacts
front view:
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
Zero effect on breast size. Additionally, strings like “flat csist: 1.5” appear as text drawn in the background — a text artifact. Weighting syntax should not be used with z-image-turbo.
Stacking / Repetition / Negative expressions — All failed
flat chest, small breasts, no cleavage (stacking, front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
flat flat flat chest (repetition, front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
flat chest, not big breasts, not busty (negative expression, front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
Stacking (3 words), repetition (flat×3), and negative expressions (not big breasts, not busty) all failed. Negative expressions may risk backfiring since big breasts and busty are recognized as tokens.
Additional Testing: Cinderella Bust, Male Chest, Pettanko (Group H — All 6 Failed)
After 29 conditions all failed, 6 additional conditions were tested with different approaches.
cinderella bust / male chest / pettanko — All failed
cinderella bust:
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
male chest:
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
pettanko:
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
cinderella bust— “Cinderella” is interpreted as a dress costume. Breast size unchangedmale chest— The female specification from1girl+actressis strong; male chest is completely ignoredpettanko— Likely not in the model’s vocabulary or not connected to breast size
man body / 1boy / delicate — All failed (with side effects)
man body, female face:
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
1boy, flat chest:
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
delicate slim chest, barely any breast:
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
man body, female face— Ignored due to conflict with1girl1boy, flat chest— Destructive side effect: the female chest is unchanged, but an additional male figure is added to the framedelicate slim chest, barely any breast— Long natural language description also has no effect; only makes the composition unstable
Summary
All 14 conditions — maturity reinforcement (mature woman, adult woman, age emphasis, occupation setting), emphasis testing (weighting, stacking, repetition, negative expressions), and additional testing (cinderella bust, male chest, pettanko, man body, 1boy, delicate) — failed.
Notable side effects to be aware of:
- Weighting syntax
(flat chest:1.5)causes text artifacts - Negative expressions
not big breastsrisk backfiring 1boyadds a separate male figure — a destructive side effect
For other approach results, see:
- Hub article: Flat Chest Prompt Test (Overall Conclusions)
- Part 1: Direct descriptions and cup size notation (Groups B+C)
- Part 2: Body type and specific descriptions (Groups D+E)
![[Verified] Image Generation Prompt Best Practices](/tips/prompt-best-practices/cover_0_0000_4517457392071889496.webp)
