Saved in:
Bibliographic Details
Main Authors: Schneider, Sinclair, Steuber, Florian, Schneider, Joao A. G., Rodosek, Gabi Dreo
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.07595
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913728347242496
author Schneider, Sinclair
Steuber, Florian
Schneider, Joao A. G.
Rodosek, Gabi Dreo
author_facet Schneider, Sinclair
Steuber, Florian
Schneider, Joao A. G.
Rodosek, Gabi Dreo
contents The increasing popularity of large language models has not only led to widespread use but has also brought various risks, including the potential for systematically spreading fake news. Consequently, the development of classification systems such as DetectGPT has become vital. These detectors are vulnerable to evasion techniques, as demonstrated in an experimental series: Systematic changes of the generative models' temperature proofed shallow learning-detectors to be the least reliable. Fine-tuning the generative model via reinforcement learning circumvented BERT-based-detectors. Finally, rephrasing led to a >90\% evasion of zero-shot-detectors like DetectGPT, although texts stayed highly similar to the original. A comparison with existing work highlights the better performance of the presented methods. Possible implications for society and further research are discussed.
format Preprint
id arxiv_https___arxiv_org_abs_2503_07595
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Detection Avoidance Techniques for Large Language Models
Schneider, Sinclair
Steuber, Florian
Schneider, Joao A. G.
Rodosek, Gabi Dreo
Computation and Language
The increasing popularity of large language models has not only led to widespread use but has also brought various risks, including the potential for systematically spreading fake news. Consequently, the development of classification systems such as DetectGPT has become vital. These detectors are vulnerable to evasion techniques, as demonstrated in an experimental series: Systematic changes of the generative models' temperature proofed shallow learning-detectors to be the least reliable. Fine-tuning the generative model via reinforcement learning circumvented BERT-based-detectors. Finally, rephrasing led to a >90\% evasion of zero-shot-detectors like DetectGPT, although texts stayed highly similar to the original. A comparison with existing work highlights the better performance of the presented methods. Possible implications for society and further research are discussed.
title Detection Avoidance Techniques for Large Language Models
topic Computation and Language
url https://arxiv.org/abs/2503.07595