AI-Generated Text Identification Feature Library
A systematic analysis of core differences between AI-generated and human-written text across linguistic, structural, emotional, and personal dimensions, providing actionable detection framework and metrics
Prompt Content
Copy and paste directly into your model or internal evaluation tool.
Analyze whether the given text is AI-generated based on the following feature library. Focus on: 1) frequent use of fixed patterns like 'A is not X but Y' or 'X and Y'; 2) over-reliance on dashes, colons, and quotation marks; 3) excessive use of explicit connectors like 'secondly', 'therefore'; 4) presence of translationese or overly academic tone; 5) flat emotional expression and lack of personalization; 6) mechanical 'general-specific-general' structure; 7) absence of viewpoint sources and expressions of uncertainty. Make a comprehensive judgment using high, medium, and low-weight indicators, and suggest reverse de-AI testing.
Use Cases
Reference Output
The text frequently uses the 'true X lies not in... but in...' structure, with dash usage density reaching 3.2 per 100 words, and shows no emotional fluctuation or personal expression throughout. These align with high-weight indicators of AI generation. Initial assessment suggests high likelihood of AI origin. Recommend conducting reverse de-AI testing for further validation.
Scoring Rubric
Scoring dimensions include: Sentence pattern rigidity (0-3 pts), Punctuation anomaly frequency (0-2 pts), Connector overuse (0-3 pts), Emotional flatness (0-2 pts), Structural predictability (0-3 pts), Lack of personalization (0-2 pts). Total score ≥8 indicates high probability of AI generation, 5-7 suggests suspicion, ≤4 indicates likely human authorship.
Try & save
Fill variables and copy, or save as a personal template.
This template has no variables and is ready to copy.
User Rating
0 ratingsYour rating
Log in to rate
Comments
0Log in to comment
Related Prompts
Agent Safety Testing at Scale Architect
Design an automated, scalable safety-testing system for LLM agents using the three-stage Vera pipeline: risk discovery, executable safety-case generation, and deterministic sandbox verification.
China Software Copyright Materials Architect
Generates complete, reviewable, and submission-ready Chinese software copyright (软件著作权) registration packages directly from a real project: application form fields, operational manuals for non-technical examiners, and code materials compliant with CNIPA rules.
Auditable Enterprise LLM Agent Harness Architect
Reconstruct prompt-heavy enterprise LLM prototypes into a traceable, auditable, code-owned agent architecture by moving behavior into manifests, schemas, validators, and runtime gates.
Agentmemory Persistent Memory Architect
Prompt from prompts: Agentmemory Persistent Memory Architect