介绍
{ "text": "# AI 文本拟人化\n\n全面的命令行工具 (CLI)，用于检测和转换 AI 生成的文本以绕过检测器。基于 [Wikipedia 的 AI 写作迹象](https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing)。\n\n## 快速开始\n\n```bash # Humanize AI Text python scripts/detect.py text.txt
# Transform to human-like python scripts/transform.py text.txt -o clean.txt
# Compare before/after python scripts/compare.py text.txt -o clean.txt ```\n\n---\n\n## 检测分类\n\n分析器会检查 Wikipedia 指南中的 **16 种模式类别**：\n\n### 严重（立即检测出 AI）\n| 类别 | 示例 |\n|----------|----------|\n| 引用错误 | `oaicite`, `turn0search`, `contentReference` |\n| 知识截止 | \"as of my last training\", \"based on available information\" |\n| 聊天机器人特征 | \"I hope this helps\", \"Great question!\", \"As an AI\" |\n| Markdown | `**bold**`, `## headers`, ``` code blocks ``` |\n\n### 高度信号\n| 类别 | 示例 |\n|----------|----------|\n| AI 词汇 | delve, tapestry, landscape, pivotal, underscore, foster |\n| 意义夸大 | \"serves as a testament\", \"pivotal moment\", \"indelible mark\" |\n| 推销语言 | vibrant, groundbreaking, nestled, breathtaking |\n| 系动词回避 | 用 \"serves as\" 代替 \"is\"，用 \"boasts\" 代替 \"has\" |\n\n### 中度信号\n| 类别 | 示例 |\n|----------|----------|\n| 表面 -ing 形式 | \"highlighting the importance\", \"fostering collaboration\" |\n| 填充短语 | \"in order to\", \"due to the fact that\", \"Additionally,\" |\n| 模糊归因 | \"experts believe\", \"industry reports suggest\" |\n| 挑战公式 | \"Despite these challenges\", \"Future outlook\" |\n\n### 风格信号\n| 类别 | 示例 |\n|----------|----------|\n| 弯引号 | 用 \"\" 代替 \"\" （ChatGPT 的特征） |\n| 破折号滥用 | 过度使用 — 来表示强调 |\n| 否定排比 | \"Not only... but also\", \"It's not just... it's\" |\n| 三法则 | 强制性的三连词，如 \"innovation, inspiration, and insight\" |\n\n---\n\n## 脚本\n\n### detect.py — 扫描 AI 模式\n\n```bash python scripts/detect.py essay.txt python scripts/detect.py essay.txt -j # JSON output python scripts/detect.py essay.txt -s # score only echo "text" | python scripts/detect.py ```\n\n**输出：**\n- 问题数量和字数\n- AI 概率（低/中/高/极高）\n- 按类别的详细细分\n- 标记出可自动修复的模式\n\n### transform.py — 重写文本\n\n```bash python scripts/transform.py essay.txt python scripts/transform.py essay.txt -o output.txt python scripts/transform.py essay.txt -a # aggressive python scripts/transform.py essay.txt -q # quiet ```\n\n**自动修复：**\n- 引用错误 (oaicite, turn0search)\n- Markdown (**, ##, ```) - Chatbot sentences - Copula avoidance → "is/has" - Filler phrases → simpler forms - Curly → straight quotes
**Aggressive (-a):** - Simplifies -ing clauses - Reduces em dashes
### compare.py — Before/After Analysis
```bash\npython scripts/compare.py essay.txt\npython scripts/compare.py essay.txt -a -o clean.txt\n```
Shows side-by-side detection scores before and after transformation
---
## Workflow
1. **Scan** for detection risk: ```bash\n python scripts/detect.py document.txt\n ```
2. **Transform** with comparison: ```bash\n python scripts/compare.py document.txt -o document_v2.txt\n ```
3. **Verify** improvement: ```bash\n python scripts/detect.py document_v2.txt -s\n ```
4. **Manual review** for AI vocabulary and promotional language (requires judgment)
---
## AI Probability Scoring
| Rating | Criteria | |--------|----------| | Very High | Citation bugs, knowledge cutoff, or chatbot artifacts present | | High | >30 issues OR >5% issue density | | Medium | >15 issues OR >2% issue density | | Low | <15 issues AND <2% density |
---
## Customizing Patterns
Edit `scripts/patterns.json` to add/modify: - `ai_vocabulary` — words to flag - `significance_inflation` — puffery phrases - `promotional_language` — marketing speak - `copula_avoidance` — phrase → replacement - `filler_replacements` — phrase → simpler form - `chatbot_artifacts` — phrases triggering sentence removal
---
## Batch Processing
```bash\n# 扫描所有文件\nfor f in *.txt; do\n echo \"=== $f ===\"\n python scripts/detect.py \"$f\" -s\ndone\n\n# 转换所有 markdown\nfor f in *.md; do\n python scripts/transform.py \"$f\" -a -o \"${f%.md}_clean.md\" -q\ndone\n