Lightweight Stylistic Consistency Profiling: Robust Detection of LLM-Generated Textual Content for Multimedia Moderation

TL;DR

LiSCP is a lightweight stylistic profiling method that robustly detects LLM-generated text by analyzing stylistic stability across paraphrased variants, outperforming existing detectors especially under adversarial conditions.

Contribution

Introduces LiSCP, a novel stylistic consistency profiling approach combining stylistic and semantic features for robust LLM-generated text detection.

Findings

LiSCP outperforms existing methods by up to 11.79% in cross-domain detection.

LiSCP demonstrates robustness against adversarial attacks and hybrid human-AI scenarios.

It achieves superior in-domain detection performance across multiple datasets.

Abstract

The increasing prevalence of Large Language Models (LLMs) in content creation has made distinguishing human-written textual content from LLM-generated counterparts a critical task for multimedia moderation. Existing detectors often rely on statistical cues or model-specific heuristics, making them vulnerable to paraphrasing and adversarial manipulations, and consequently limiting their robustness and interpretability. In this work, we proposeLiSCP , a novel lightweight stylistic consistency profiling method for robust detection of LLM-generated textual content, focusing on feature stability under adversarial manipulation. Our approach constructs a consistency profile that combines discrete stylistic features with continuous semantic signals, leveraging stylistic stability across multimodal-guided paraphrased text variants. Experiments spanning real-world multimedia news and movie datasets and conventional text domains demonstrate that LiSCP achieves superior performance on in-domain detection and outperforms existing approaches by up to 11.79% in cross-domain settings. Additionally,it demonstrates notable robustness under adversarial scenarios, including adversarial attacks and hybrid human-AI settings.