Efficient Black-Box Adversarial Attacks on Neural Text Detectors

Vitalii Fishchuk, Daniel Braun
11/3/2023
cs.CL

No confident code implementation found for this paper. 2 possible matches can be shown below.

Abstract

Neural text detectors are models trained to detect whether a given text was generated by a language model or written by a human. In this paper, we investigate three simple and resource-efficient strategies (parameter tweaking, prompt engineering, and character-level mutations) to alter texts generated by GPT-3.5 that are unsuspicious or unnoticeable for humans but cause misclassification by neural text detectors. The results show that especially parameter tweaking and character-level mutations are effective strategies.

View on arXivView PDF

Code Implementations

No confident code match yet

We couldn't find an author-owned or strongly-evidenced community implementation for this paper. 2 weaker matches are hidden by default — verify before relying on them.

No code implementations found yet.

Know of an implementation? Let us know in the comments below!

Similar Papers

Evade ChatGPT Detectors via A Single Space

90% match2023

Smaller Language Models are Better Black-box Machine-Generated Text Detectors

88% match2023

GPT-Sentinel: Distinguishing Human and ChatGPT Generated Content

88% match1 repo92023

Automatic detection of Gen-AI texts: A comparative framework of neural models

87% match2026

Base Models Look Human To AI Detectors

87% match2026

Ghostbuster: Detecting Text Ghostwritten by Large Language Models

87% match2 repos1872023

Cite this paper

@article{fishchuk2023efficient,
  title  = {Efficient Black-Box Adversarial Attacks on Neural Text Detectors},
  author = {Vitalii Fishchuk and Daniel Braun},
  year   = {2023},
  eprint = {2311.01873},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2311.01873v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2311.01873.svg)](https://codeofpaper.com/paper/2311.01873)

Discussion