Prompt 正文
默认展示英文原文。点击“用它生成”后,首页会默认载入此草稿。
A smartphone screenshot of an AI chat app interface in portrait orientation, showing a clean white iOS-style screen with a top bar containing a hamburger menu on the left, the blue title “Thinking” with a small chevron, and two icons on the right for compose and more options. Near the top is a rounded light-gray user message bubble in Japanese reading “Attention機構を人体に喩えて細かくラベルを付けて。医学書の精緻な挿絵風に。” Below it, faint gray system text reads “Thought for 5s” and “画像が作成されました”. The lower half of the screen displays the generated image as a detailed medical textbook-style anatomical poster. The poster shows a realistic three-quarter back view of an adult woman from head to upper thighs, with her head turned slightly left and her dark brown hair tied into a messy bun. Her skin and outer tissues are removed across most of the body to reveal musculature, nerves, blood vessels, and fibrous structures, rendered in extremely detailed vintage medical illustration style on a warm cream paper background. A rectangular cutout obscures the center of the face from forehead to jaw, creating a censored blank block over the facial area. Fine glowing network lines and nodes are overlaid through the exposed brain and body to visualize the attention mechanism as if it were an anatomical system. Surrounding the figure are 6 numbered explanatory panels and 2 additional text blocks, all densely filled with Japanese and English bilingual labels, formulas, arrows, and diagrams. The main poster title at upper left is “アテンション機構の解剖学的構造” with the English subtitle “The Anatomy of the Attention Mechanism”. At the top right is a flowchart labeled “Overall Flow of Attention” containing 5 connected boxes describing the process from input to output. Down the left side are 3 stacked explanatory sections with equations and mapping tables. Down the right side are 3 more numbered sections, each with technical notes and a small chart or matrix; the visible mini-diagrams include 1 purple heatmap matrix, 1 blue bar chart, and 1 multicolored cube graphic. Along the bottom edge are 5 small brain inset illustrations with captions showing stages of information flow. The overall aesthetic should feel like a highly polished AI-generated screenshot documenting a surreal scientific illustration request, with crisp UI, lots of tiny readable labels, and a convincing fusion of transformer attention diagrams and classical anatomical plate design.
猜你喜欢

3D 石阶演化信息图
将平面的演化时间轴转化为逼真的 3D 石阶信息图,包含精细的生物渲染图和结构化的侧边栏。

信息图 / 教育视觉图 - 糕点城市建筑艺术品
2x2 网格,16:9,制作麦加及其他 3 个著名城市,并使用 4 种不同的渲染引擎=> 一款甜点风格的艺术品,主体为 {argument name="city" default="MECCA"};采用法式甜酥面团墙壁,切面可见层压纹理,撒...

信息图 / 教育视觉图 - 儿童马克笔画风格
以用户输入图像的具体主体内容为中心,将其重绘为一张横向白纸上的儿童马克笔手绘图。首先,从输入内容中提取最具辨识度的外轮廓、姿态方向、局部地标、功能组件及身份色彩,将复杂的细节压缩为几个清晰的大形状;主体不追求写实的比例或精确的结构,而是通过...

信息图 / 教育视觉图 - 建筑蓝图肖像
超精细肖像,采用现代主义插画风格,灵感源自建筑蓝图和技术绘图。面部被分割为多个几何图层,填充有电路图、城市示意图、建筑网格和结构蓝图。背景为清爽的白色,采用浓密的黑色墨水线条,辅以交叉排线和雕刻阴影。深红色、钢蓝色和琥珀黄作为平涂色块和半透...
