MGTeval : An Interactive Platform for Systemtic Evaluation of Machine-Generated Text Detectors

Yuanfan Li, Qi Zhou, Chengzhengxu Li, Zhaohan Zhang, Chenxu Zhao, Zepu Ruan, Chao Shen, Xiaoming Liu · 2026

Read MGTeval : An Interactive Platform for Systemtic Evaluation of Machine-Generated Text Detectors by Yuanfan Li, Qi Zhou, Chengzhengxu Li, Zhaohan Zhang, Chenxu Zhao, Zepu Ruan, Chao Shen, Xiaoming Liu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-25152.txt.

Description

Li and seven colleagues present MGTEVAL, a platform for systematic evaluation of machine-generated text detectors. Current evaluations scatter across data sets, preprocessing, attacks and metrics, so results are hard to compare or reproduce. MGTEVAL splits the work into data set building, attack, detector training and evaluation. It builds custom benchmarks with configurable language models, applies 12 text attacks to test sets and reports effectiveness, robustness and efficiency via command-line and web interfaces.

Catalog record

CREATORYuanfan Li, Qi Zhou, Chengzhengxu Li, Zhaohan Zhang, Chenxu Zhao, Zepu Ruan, Chao Shen, Xiaoming Liu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE30675
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.25152

Browse the full library · About the Tower of Babel