Unmasking the Imposters: How Censorship and Domain Adaptation Affect the Detection of Machine-Generated Tweets

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Tuck, Bryan E., Verma, Rakesh M.
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929682722586624
author Tuck, Bryan E.
Verma, Rakesh M.
author_facet Tuck, Bryan E.
Verma, Rakesh M.
contents The rapid development of large language models (LLMs) has significantly improved the generation of fluent and convincing text, raising concerns about their potential misuse on social media platforms. We present a comprehensive methodology for creating nine Twitter datasets to examine the generative capabilities of four prominent LLMs: Llama 3, Mistral, Qwen2, and GPT4o. These datasets encompass four censored and five uncensored model configurations, including 7B and 8B parameter base-instruction models of the three open-source LLMs. Additionally, we perform a data quality analysis to assess the characteristics of textual outputs from human, "censored," and "uncensored" models, employing semantic meaning, lexical richness, structural patterns, content characteristics, and detector performance metrics to identify differences and similarities. Our evaluation demonstrates that "uncensored" models significantly undermine the effectiveness of automated detection methods. This study addresses a critical gap by exploring smaller open-source models and the ramifications of "uncensoring," providing valuable insights into how domain adaptation and content moderation strategies influence both the detectability and structural characteristics of machine-generated text.
format Preprint
id arxiv_https___arxiv_org_abs_2406_17967
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Unmasking the Imposters: How Censorship and Domain Adaptation Affect the Detection of Machine-Generated Tweets
Tuck, Bryan E.
Verma, Rakesh M.
Computation and Language
The rapid development of large language models (LLMs) has significantly improved the generation of fluent and convincing text, raising concerns about their potential misuse on social media platforms. We present a comprehensive methodology for creating nine Twitter datasets to examine the generative capabilities of four prominent LLMs: Llama 3, Mistral, Qwen2, and GPT4o. These datasets encompass four censored and five uncensored model configurations, including 7B and 8B parameter base-instruction models of the three open-source LLMs. Additionally, we perform a data quality analysis to assess the characteristics of textual outputs from human, "censored," and "uncensored" models, employing semantic meaning, lexical richness, structural patterns, content characteristics, and detector performance metrics to identify differences and similarities. Our evaluation demonstrates that "uncensored" models significantly undermine the effectiveness of automated detection methods. This study addresses a critical gap by exploring smaller open-source models and the ramifications of "uncensoring," providing valuable insights into how domain adaptation and content moderation strategies influence both the detectability and structural characteristics of machine-generated text.
title Unmasking the Imposters: How Censorship and Domain Adaptation Affect the Detection of Machine-Generated Tweets
topic Computation and Language
url https://arxiv.org/abs/2406.17967