圖像 提示詞
洩漏的 AI 基準測試報告照片
生成一張逼真的電腦螢幕照片，螢幕上顯示著包含長條圖與詳細效能表格的學術技術報告。
提示詞
{
"type": "顯示學術技術報告的電腦螢幕照片",
"style": "略帶角度的螢幕拍攝視角，可見摩爾紋、LCD 像素網格、輕微反光、LaTeX 文件格式、襯線字體",
"document_header": {
"left": "4 基準測試評估",
"right": "DeepSeek-V4 技術報告"
},
"introductory_text": "總結 DeepSeek-V4 對比 GPT-5.3、Claude Opus 4.6 以及 Gemini 3.1 Pro Preview 之綜合評估的段落。",
"visualizations": {
"legend": "5 個項目及其顏色代碼：深藍色、灰色、淺灰色、藍色條紋、淺藍色",
"bar_charts": {
"count": 6,
"labels": [
"MMLU-Pro (EM)",
"GPQA-Diamond (Pass@1)",
"AIME 2025 (Pass@1)",
"LiveCodeBench (Pass@1-COT)",
"SWE-bench Verified (Resolved)",
"Tau-bench (Average)"
]
},
"caption": "圖 1 | 核心基準測試效能比較。DeepSeek-V4 在大多數基準測試中均達到業界領先水準。"
},
"data_table": {
"columns": [
"基準測試",
"DeepSeek-V4",
"GPT-5.3",
"Claude Opus 4.6",
"Gemini 3.1 Pro Preview",
"GPT-4.1"
],
"categories": {
"count": 4,
"rows": [
{"label": "通用", "icon": "globe/network", "sub_items": 3},
{"label": "推理與數學", "icon": "calculator/clipboard", "sub_items": 3},
{"label": "程式碼", "icon": "code brackets", "sub_items": 3},
{"label": "Agent", "icon": "robot face", "sub_items": 3}
]
}
}
}
