Connect

Chinese AI Content Humanizer

brycewang-stanfordbybrycewang-stanford
•
1.9k
•
Content Management

Detects and transforms AI-generated Chinese text into natural, human-like prose using advanced statistical modeling and style-specific rewrites.

This skill provides a comprehensive toolset for identifying and mitigating 'AI-flavored' Chinese text across various domains including academic research, creative writing, and social media. By leveraging scene-aware Logistic Regression fusion and over 20 rule-based detection categories, it helps users bypass AIGC detectors (like CNKI or Wanfang) while enhancing text readability through style transformations (e.g., Zhihu, Xiaohongshu, Academic). It is particularly effective for researchers and content creators who need to maintain linguistic naturalness and satisfy rigorous AI-detection criteria in high-stakes environments.

Key Features

01Best-of-N humanization with statistical optimization (up to N=10)
02AIGC reduction optimized for CNKI, VIP, and Wanfang databases
031,853 GitHub stars
04Scene-aware detection models (General, Academic, Long-form)
05Comprehensive linguistic analysis (Perplexity, GLTR, and sentence-length CV)
068 distinct style transformations including Xiaohongshu and Zhihu

Use Cases

01Transforming AI-generated marketing copy into authentic social media posts
02Reducing AIGC scores in academic papers for formal journal submissions
03Humanizing long-form novels and blogs to remove repetitive AI artifacts