Can You Detect the Difference?

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Tarım, İsmail, Onan, Aytuğ
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915389215080448
author Tarım, İsmail
Onan, Aytuğ
author_facet Tarım, İsmail
Onan, Aytuğ
contents The rapid advancement of large language models (LLMs) has raised concerns about reliably detecting AI-generated text. Stylometric metrics work well on autoregressive (AR) outputs, but their effectiveness on diffusion-based models is unknown. We present the first systematic comparison of diffusion-generated text (LLaDA) and AR-generated text (LLaMA) using 2 000 samples. Perplexity, burstiness, lexical diversity, readability, and BLEU/ROUGE scores show that LLaDA closely mimics human text in perplexity and burstiness, yielding high false-negative rates for AR-oriented detectors. LLaMA shows much lower perplexity but reduced lexical fidelity. Relying on any single metric fails to separate diffusion outputs from human writing. We highlight the need for diffusion-aware detectors and outline directions such as hybrid models, diffusion-specific stylometric signatures, and robust watermarking.
format Preprint
id arxiv_https___arxiv_org_abs_2507_10475
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Can You Detect the Difference?
Tarım, İsmail
Onan, Aytuğ
Computation and Language
Artificial Intelligence
I.2.7; H.3.3
The rapid advancement of large language models (LLMs) has raised concerns about reliably detecting AI-generated text. Stylometric metrics work well on autoregressive (AR) outputs, but their effectiveness on diffusion-based models is unknown. We present the first systematic comparison of diffusion-generated text (LLaDA) and AR-generated text (LLaMA) using 2 000 samples. Perplexity, burstiness, lexical diversity, readability, and BLEU/ROUGE scores show that LLaDA closely mimics human text in perplexity and burstiness, yielding high false-negative rates for AR-oriented detectors. LLaMA shows much lower perplexity but reduced lexical fidelity. Relying on any single metric fails to separate diffusion outputs from human writing. We highlight the need for diffusion-aware detectors and outline directions such as hybrid models, diffusion-specific stylometric signatures, and robust watermarking.
title Can You Detect the Difference?
topic Computation and Language
Artificial Intelligence
I.2.7; H.3.3
url https://arxiv.org/abs/2507.10475