先重写结构，再谈降 AI 率 - 论文辅助

原文要点

Computer Science > Artificial Intelligence arXiv:2506.23517 (cs) [Submitted on 30 Jun 2025] Title: Assessing GPTZero's Accuracy in Identifying AI vs. Human-Written Essays Authors: Selin Dik , Osman Erdem , Mehmet Dik View a PDF of the paper titled Assessing GPTZero's Accuracy in Identifying AI vs. Human-Written Essays, by Selin Dik and 1 other authors View PDF Abstract: As the use of AI tools by students has become more prevalent, instructors have started using AI detection tools like GPTZero and QuillBot to detect AI written text. However, the reliability of these detectors remains uncertain. In our study, we focused mostly on the success rate of GPTZero, the most-used AI detector, in identifying AI-generated texts based on different lengths of randomly submitted essays: short (40-100 word count), medium (100-350 word count), and long (350-800 word count). We gathered a data set consisting of twenty-eight AI-generated papers and fifty human-written papers. With this randomized essay d

原文完整内容

Computer Science > Artificial Intelligence

arXiv:2506.23517

(cs)

[Submitted on 30 Jun 2025]

Title:

Assessing GPTZero's Accuracy in Identifying AI vs. Human-Written Essays

Authors:

Selin Dik

Osman Erdem

Mehmet Dik

View a PDF of the paper titled Assessing GPTZero's Accuracy in Identifying AI vs. Human-Written Essays, by Selin Dik and 1 other authors

View PDF

Abstract:

As the use of AI tools by students has become more prevalent, instructors have started using AI detection tools like GPTZero and QuillBot to detect AI written text. However, the reliability of these detectors remains uncertain. In our study, we focused mostly on the success rate of GPTZero, the most-used AI detector, in identifying AI-generated texts based on different lengths of randomly submitted essays: short (40-100 word count), medium (100-350 word count), and long (350-800 word count). We gathered a data set consisting of twenty-eight AI-generated papers and fifty human-written papers. With this randomized essay data, papers were individually plugged into GPTZero and measured for percentage of AI generation and confidence. A vast majority of the AI-generated papers were detected accurately (ranging from 91-100% AI believed generation), while the human generated essays fluctuated; there were a handful of false positives. These findings suggest that although GPTZero is effective at detecting purely AI-generated content, its reliability in distinguishing human-authored texts is limited. Educators should therefore exercise caution when relying solely on AI detection tools.

Subjects:

Artificial Intelligence (cs.AI)

; Computation and Language (cs.CL)

Cite as:

arXiv:2506.23517

[cs.AI]

(or

arXiv:2506.23517v1

原文配图

六、把结构重写落到实处

每段写一句中心句。
中心句下面只放支撑句，不放跑题句。
段落之间建立“问题 - 方法 - 结果 - 限制”的关系。
删除与中心无关的解释。
确保每一段都有明确信息增量。

七、改写前后对比

示例

原句：
It is worth noting that the proposed method achieves better performance
due to various reasons.

改为：
We compared the proposed method with baseline A on two datasets.
The gain came mainly from the contrastive loss, not from extra parameters.

来源：https://arxiv.org/abs/2506.23517