![[Experiment] Flat Chest Prompts (Part 1) | Direct Descriptions and Cup Size Notation](/tips/flat-chest-descriptor-test-part1/a00_control_front_s1_0_0000_2053695854357871006.webp)
Conclusions
- All 6 direct description conditions failed —
flat chest,small breasts,tiny breasts,no breasts,completely flat chest, andflat chestedwere all visually indistinguishable from control - All 3 cup size notation conditions also failed —
A-cup,AA-cup, andmicrohad no effect. Separate testing confirmed that going larger (G-cup, etc.) also doesn’t work large breastsstably works in 3/3 to increase size — confirming an asymmetry: only the “make bigger” direction worksno breaststriggers a side effect of compositions that “hide” the chest — interpreted as “don’t show breasts” rather than “have no breasts”
This article is Part 1 of the hub article on flat chest / small chest prompt testing. It covers 9 conditions (54 images) of direct size descriptions (Group B) and cup size notation (Group C).
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.
Direct Descriptions (Group B — All 6 Failed)
The most intuitive descriptions were tried. All failed.
flat chest — No effect
front view:
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
side view:
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
Visually indistinguishable from control. “flat” occasionally causes butterfly or floral decorations to appear in the background as a side effect.
small breasts / tiny breasts / no breasts — All failed
small breasts (front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
tiny breasts (front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
no breasts (front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
completely flat chest (front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
small, tiny, no, completely flat — even increasing the expression strength step by step had zero effect on size. no breasts triggers a side effect where cloth or props “hide” the chest (interpreted as “don’t show breasts” rather than “have no breasts”).
Cup Size Notation (Group C — All 3 Failed)
A-cup / AA-cup / micro — All failed
A-cup breasts (front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
AA-cup breasts (front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
micro breasts (front):
| seed 1 | seed 2 | seed 3 |
|---|---|---|
 |  |  |
Cup size notation for larger sizes (G-cup, etc.) was also confirmed to be ineffective in the breast size cup notation test. The same applies for smaller sizes.
Summary
All 9 conditions — direct descriptions (flat chest, small breasts, tiny breasts, no breasts, completely flat chest, flat chested) and cup size notation (A-cup, AA-cup, micro) — failed. The most intuitive prompts that work on other models are completely ineffective here.
For other approach results, see:
- Hub article: Flat Chest Prompt Test (Overall Conclusions)
- Part 2: Body type and specific descriptions (Groups D+E)
- Part 3: Maturity reinforcement, emphasis, additional testing (Groups F+G+H)
![[Verified] Image Generation Prompt Best Practices](/tips/prompt-best-practices/cover_0_0000_4517457392071889496.webp)
