diff --git a/README.md b/README.md
index a42f356..3b4f809 100644
--- a/README.md
+++ b/README.md
@@ -92,7 +92,7 @@ Both aim to restore diversity in distilled models, but they work at fundamentall
 | **Safety** | Amplitude exactly preserved; Butterworth + tanh double-bounded | No mathematical guarantees |
 | **Prompt adherence** | Unaffected (operates on spatial structure, not semantics) | May degrade (alters attention distribution) |
 
-In short: DiversityBoost is **precise spatial surgery** — it changes where things are placed without affecting what they look like. Dummy token is **injecting noise into the model's input** hoping for a butterfly effect.
+In short: DiversityBoost operates precisely on composition-scale phase in the frequency domain, with mathematically bounded safety guarantees. Dummy token injects perturbation at the token level — the effect is indirect and unpredictable.
 
 ## Tips
 
diff --git a/README_zh.md b/README_zh.md
index 91a2f76..60f1318 100644
--- a/README_zh.md
+++ b/README_zh.md
@@ -92,7 +92,7 @@ MODEL → [Diversity Boost] → MODEL → KSampler
 | **安全性** | 振幅精确保持；Butterworth + tanh 双重限幅 | 无数学保证 |
 | **prompt 遵循度** | 不受影响（操作空间结构，不涉及语义） | 可能下降（改变 attention 分布） |
 
-简单说：DiversityBoost 是**精确的空间手术**——改变物体的位置布局，不影响物体本身的样貌。Dummy token 是**往模型输入里注入噪声**，依赖蝴蝶效应产生不同结果。
+简单说：DiversityBoost 在频域精确操作构图尺度的相位，有数学边界保证安全性。Dummy token 在 token 层面注入扰动，效果间接且不可预测。
 
 ## 使用建议