NLP-Based Review for Toxic Comment Detection Tailored to the Chinese Cyberspace

TL;DR

The paper addresses NLP-based toxic comment detection in the Chinese cyberspace, focusing on platform-specific language, rapid semantic evolution, and cultural nuances. It surveys toxicity definitions, platform ecosystems, and dataset construction, then outlines the evolution of detection models from rules to deep learning and LLMs, with emphasis on interpretability. A novel fine-grained toxicity framework and multi-dimensional annotation strategy are proposed, along with efficient human–machine data curation and robust quality assurance. The work also identifies open problems—contextual heterogeneity, multimodal signals, and data noise—and offers concrete directions such as platform-aware taxonomy, adversarial robustness, and multimodal fusion for practical, scalable governance of online toxicity in China.

Abstract

With the in-depth integration of mobile Internet and widespread adoption of social platforms, user-generated content in the Chinese cyberspace has witnessed explosive growth. Among this content, the proliferation of toxic comments poses severe challenges to individual mental health, community atmosphere and social trust. Owing to the strong context dependence, cultural specificity and rapid evolution of Chinese cyber language, toxic expressions are often conveyed through complex forms such as homophones and metaphors, imposing notable limitations on traditional detection methods. To address this issue, this review focuses on the core topic of natural language processing based toxic comment detection in the Chinese cyberspace, systematically collating and critically analyzing the research progress and key challenges in this field. This review first defines the connotation and characteristics of Chinese toxic comments, and analyzes the platform ecology and transmission mechanisms they rely on. It then comprehensively reviews the construction methods and limitations of existing public datasets, and proposes a novel fine-grained and scalable framework for toxic comment definition and classification, along with corresponding data annotation and quality assessment strategies. We systematically summarize the evolutionary path of detection models from traditional methods to deep learning, with special emphasis on the importance of interpretability in model design. Finally, we thoroughly discuss the open challenges faced by current research and provide forward-looking suggestions for future research directions.

Figures (6)

• Figure 1: Structure of the Review and Topics Discussed in Each Section.
• Figure 2: Conceptual Hierarchy of Toxic Comment.
• Figure 3: Closed-Loop Workflow of Human-Machine Collaborative Annotation.
• Figure 4: Schematic diagram of the LIME technical workflow.
• Figure 5: Attention Weight Heatmap of Chinese Toxic Comments Based on the BiLSTM-Attention Model.