Conclusion
- Metaphors generate unintended objects —
hanging bombsproduces explosives in the background,watermelon-sizedproduces watermelons,melonproduces melons. None contribute to breast size. - Japanese romaji (oppai) and dairy cow expressions (holstein) have no effect — holstein produces a Holstein cow in every pattern.
super/strongprefixes alone have no size-increasing effect — The full stack’s effectiveness comes from the total volume of size-related tokens.- Replacing
tits/boobswithbreastsdoesn’t prevent nipple exposure — N=13, all had nipple exposure. busty,buxom,breast expansionare limited — aroundlargelevel — No significant size increase compared to confirmed effective descriptions (huge, massive, enormous, oversized).
This is the detailed article (Part 2) for Comprehensive Breast Size Prompt Comparison. It covers slang, metaphors, and syntax analysis approaches.
- Part 1 (cup sizes, cm measurements, weight syntax): here
- Effective descriptions: see this article
Experiment conditions and base prompt: see the summary article.
Metaphors and Slang
hanging bombs — Explosives appear in the background
![]() | ![]() | ![]() |
titty monster — Breast size comparable to control
![]() | ![]() | ![]() |
watermelon-sized breasts — Image becomes someone holding a watermelon
![]() | ![]() | ![]() |
gigantic jugs — Stays around large level
![]() | ![]() | ![]() |
heavy breasts — Control level
![]() | ![]() | ![]() |
big tits — Around large level
![]() | ![]() | ![]() |
melon breasts — A melon appears
![]() | ![]() | ![]() |
bigger than head breasts — Control level
![]() | ![]() | ![]() |
Metaphors fail to produce the intended effect. bombs generates actual explosives; watermelon-sized produces a watermelon; melon produces a melon. big tits and heavy breasts stay around large level.
Japanese Romaji and Dairy Cow Expressions
oppai
![]() | ![]() | ![]() |
holstein breasts — Holstein cows appear in the background
![]() | ![]() | ![]() |
holstein (standalone) — Only a cow appears
![]() | ![]() | ![]() |
holstein massive tits — Cow appears, no breast size effect
![]() | ![]() | ![]() |
cow udders — Cow udders appear literally
![]() | ![]() | ![]() |
oppai produces control-level results. All holstein patterns generated a Holstein cow in the background with no effect on breast size. cow udders was similarly interpreted literally. No breast size increase from oppai or animal metaphors was observed.
Individual Effect of super / strong Prefixes
Testing each component of super strong super huge gigantic breasts separately.
strong breasts
![]() | ![]() | ![]() |
super strong breasts
![]() | ![]() | ![]() |
strong huge breasts
![]() | ![]() | ![]() |
super strong huge breasts
![]() | ![]() | ![]() |
super gigantic breasts
![]() | ![]() | ![]() |
strong alone is control level; super strong is also control level (with some seeds splitting into two people). strong huge and super strong huge show the huge effect, but no additive boost from strong/super strong. strong and super themselves have no size-increasing effect. super strong super huge gigantic breasts was large because of the total volume of size-related tokens (huge, gigantic), not because of any special power in the super strong super prefix (see size token volume hypothesis in the effective article for details).
tits/boobs → breasts Unification Test
The hypothesis: “Since tits and boobs are strongly associated with nude content, replacing them with breasts might prevent nipple rendering.”
3-seed comparison
![]() | ![]() | ![]() |
10 independent random seeds
![]() | ![]() | ![]() |
![]() | ![]() | ![]() |
![]() | ![]() | ![]() |
N=13 (3+10) — all had nipple exposure. Replacing tits/boobs with breasts does not prevent nipple exposure.
Other Limited-Effect Descriptions
busty — Around large (1-B) level
![]() | ![]() | ![]() |
buxom — Around large (1-B) level
![]() | ![]() | ![]() |
sagging nipples — Around large level. Shape descriptor, not a size booster
![]() | ![]() | ![]() |
breast expansion — Around large level. Limited effect on realistic models
![]() | ![]() | ![]() |
busty and buxom mean “large-breasted” but stay at large level. sagging nipples and breast expansion are also around large level — underwhelming.
Summary
| Method | Result |
|---|---|
| Metaphors (hanging bombs, watermelon-sized, etc.) | No effect (unintended objects appear) |
| Slang (titty monster, gigantic jugs) | No effect to large level |
| Japanese romaji (oppai) | No effect |
| Dairy cow expressions (holstein) | No effect (cow appears) |
| super/strong prefix alone | No effect |
| breasts unification (replacing tits/boobs) | Doesn’t prevent nipple exposure |
| busty, buxom, sagging nipples, breast expansion | Limited to large level |
Slang and metaphors only cause unintended objects to appear — they don’t contribute to breast size increases.
Back to summary | Effective descriptions | Part 1: Cup sizes, cm measurements, weight syntax















































































![[Verified] Image Generation Prompt Best Practices](/tips/prompt-best-practices/cover_0_0000_4517457392071889496.webp)
