【AI繪圖教學】Stable Diffusion萬用負面關鍵字（Negative Prompt）完整懶人包！防止手指變形、肢體崩壞與NSFW過濾咒語整理

【文章摘要】 在 Stable Diffusion AI 繪圖中，「正向提示詞（Positive Prompt）」負責告訴 AI 要畫出哪些元素，而「負向提示詞（Negative Prompt／負面關鍵字）」則負責明確排除不希望出現的畫面瑕疵。本篇完整整理玩家必備的「完整版防崩壞萬用負向關鍵字」與「精簡版 EasyNegative 職場安全過濾咒語」，有效防止 AI 畫出多指、斷手斷腳、五官扭曲或低畫質雜訊，並附上搭配 RION LoRA 的「圖書館看書女孩」實測對照成果！

為什麼 Stable Diffusion 必設「負向提示詞（Negative Prompt）」？

在 Stable Diffusion 的文生圖（txt2img）與圖生圖（img2img）流程中，正向關鍵字（Positive Prompt）負責告訴 AI 模型「畫面裡要畫出哪些東西」，而負向關鍵字（Negative Prompt，又稱反向提示詞或負面咒語）則是明確指示 AI「絕對不要在畫面中畫出哪些東西」。

許多剛接觸 AI 繪圖的新手常遇到人物出現「六根手指、手掌融合、肢體多出一截、五官比例崩壞」或畫面帶有模糊浮水印等問題。此時只要在 Negative Prompt 欄位填入經過社群驗證的萬用負向關鍵字，就能大幅壓制擴散模型在手部與人體骨架上的生成錯誤率，防止畫出斷手斷腳的崩壞圖！

範本一：完整版「防肢體崩壞與高畫質鎖定」萬用負向關鍵字

以下這串是我整理收集的超完整萬用負向關鍵字清單（已剔除重複冗餘詞條），涵蓋了骨架畸形（bad anatomy）、手部與手指多指變形（mutated hands and fingers:1.5、one hand with more than 5 fingers）、臉部五官崩壞、文字浮水印與低畫質雜訊排除，可直接整段複製貼入 Negative Prompt 欄位：

(((deformed))), blurry, bad anatomy, disfigured, poorly drawn face, mutation, mutated, (extra_limb), (ugly), (poorly drawn hands), fused fingers, messy drawing, broken legs, censor, censored, censor_bar, multiple breasts, (mutated hands and fingers:1.5), (long body:1.3), (mutation, poorly drawn:1.2), black-white, liquid body, liquid tongue, malformed, anatomical nonsense, text font ui, error, malformed hands, long neck, blurred, lowers, low res, bad proportions, bad shadow, uncoordinated body, unnatural body, fused breasts, bad breasts, huge breasts, poorly drawn breasts, extra breasts, liquid breasts, heavy breasts, missing breasts, huge haunch, huge thighs, huge calf, bad hands, fused hand, missing hand, disappearing arms, disappearing thigh, disappearing calf, disappearing legs, fused ears, bad ears, poorly drawn ears, extra ears, liquid ears, heavy ears, missing ears, fused animal ears, bad animal ears, poorly drawn animal ears, extra animal ears, text, ui, missing fingers, missing limb, one hand with more than 5 fingers, one hand with less than 5 fingers, one hand with more than 5 digit, one hand with less than 5 digit, extra digit, fewer digits, fused digit, missing digit, bad digit, liquid digit, colorful tongue, black tongue, cropped, watermark, username, JPEG artifacts, signature, 3D, 3D game, 3D game scene, 3D character, malformed feet, extra feet, bad feet, poorly drawn feet, fused feet, missing feet, extra shoes, bad shoes, fused shoes, more than two shoes, poorly drawn shoes, bad gloves, poorly drawn gloves, fused gloves, bad hairs, poorly drawn hairs, fused hairs, big muscles, bad face, fused face, cloned face, big face, long face, bad eyes, fused eyes, poorly drawn eyes, extra eyes, malformed limbs, gross proportions, short arm, (((missing arms))), missing thighs, missing calf, missing legs, duplicate, morbid, mutilated, more than 1 left hand, more than 1 right hand, extra arms, extra thighs, more than 2 thighs, extra calf, fused calf, extra legs, bad knee, extra knee, more than 2 legs, bad tails, bad mouth, fused mouth, poorly drawn mouth, bad tongue, too long tongue, big mouth, cracked mouth, dirty face, dirty teeth, yellow teeth, thick lips, bad collarbone, fused collarbone, missing collarbone, liquid collarbone, obesity, worst quality, low quality, normal quality, QR code, bar code, safety panties, safety knickers, beard, furry, pony, mosaic

範本二：精簡版「EasyNegative ＋ 手部修正」萬用負向關鍵字

除了長篇幅的傳統負向關鍵字外，實務上也非常推薦搭配負向 Textual Inversion 嵌入模型（如 EasyNegative 與 bad-hands-5），並自行微調不希望出現在畫面中的元素（例如排除黑白單色、皮膚斑點，或避免在工作場合不適宜出現的過度裸露 NSFW 元素）：

(EasyNegative:1.4), (worst quality:1.4), (low quality:1.4), (normal quality:1.4), bad-hands-5, lowers, monochrome, grayscales, skin spots, acnes, skin blemishes, age spot, 6 more fingers on one hand, deformity, bad legs, error legs, bad feet, malformed limbs, extra limbs

實戰生成展示：「/在圖書館看書的女孩」搭配 RION LoRA 成品

以下這張實測作品是以「在圖書館看書的女孩」為主題，套用上述萬用負向關鍵字排除手部與肢體崩壞，並搭配 RION LoRA 模型所生成的精緻畫面：

本張作品搭配之正向提示詞（Positive Prompt）：

1girl, solo, digital painting, sitting, glasses, book, library, long hair, brown hair, loose hair, hair between eyes, peaceful expression, casual clothing, indoor, shelves, table, chair, natural light, calm atmosphere, bookish.

更多 Stable Diffusion AI 繪圖系列筆記

• 環境安裝：藉由 Google Colab 來安裝 Stable Diffusion 玩 AI 繪圖
• 基本功能 txt2img（正向提示詞）：AI 繪圖筆記：使用 ChatGPT 協助產生 Stable Diffusion 正向提示詞 Positive Prompt
• 基本功能 txt2img（負向提示詞）：Stable Diffusion AI 繪圖：萬用負面關鍵字 Negative Prompt 收集

---

推薦影音清單：

以下是企業級 AI 應用與私有雲算力建置精選影音，帶您掌握從 Stable Diffusion 提示詞工程到本地端 AI 運算的完整實戰技巧！

---

相關即時資訊與推薦閱讀：

• [自由時報] 好用工具＋指令教學 從０開始一次學會AI生圖 – 整理主流AI生圖工具操作與正負向提示詞（Prompt）設定心法，協助新手快速提升生成圖片品質。
• [自由時報] 假美女裸照竟有人買！AI生成圖片網路瘋傳 4招辨真偽 – 剖析AI生成人像常見的手指關節異常、光影破綻與背景失真特徵，這也正是Negative Prompt極力修正的重點。
• [自由時報] 吳淡如遭炎上！他列舉10點知識「用AI畫圖的人叫咒術師」 – 科普AI算圖社群中的「詠唱師／咒術師」文化，說明精準調校正向與負向提示詞對畫面成敗的影響。
• [Adventure Book] 【AI繪圖教學】使用ChatGPT自動產生Stable Diffusion正向提示詞（Positive Prompt）！5大構圖要素與8K高畫質萬用咒語範本 – 教你把ChatGPT變成專屬英文提示詞產生器，一鍵產出包含5大構圖要素的8K高畫質正向咒語。
• [Adventure Book] AI繪圖筆記 藉由Google Colab 來安裝Stable Diffusion 玩AI繪圖 – 免買昂貴獨立顯卡！手把手教你透過Google Colab雲端免費GPU環境一鍵啟動Stable Diffusion WebUI。