arXiv：cs.CL· Xuchen Ma, Jianxiang Yu, Wenming Shao, Bo Pang, Xiang Li·· 11 小时前
C²TU：用同音字图谱与毒性词典揭穿中文谐音伪装毒性内容
Breaking the Cloak! Unveiling Chinese Cloaked Toxicity with Homophone Graph and Toxic Lexicon
AI 导读
研究者提出 C²TU，一种无需训练、无需提示词的方法，用于识别中文谐音伪装的有毒内容。该方法先用同音字图谱与毒性词典做子串匹配找出候选词，再过滤非毒性词并还原谐音伪装，分别基于 BERT 和 LLM 开发了两种过滤模型变体。在两个中文毒性数据集上，其 F1 最高超出最佳对比方法 71%，准确率高出 35%，代码与数据已公开。
正文
Abstract:Social media platforms have experienced a significant rise in toxic content, including abusive language and discriminatory remarks, presenting growing challenges for content moderation. Some users evade censorship by deliberately disguising toxic words through homophonic cloak, which necessitates the task of unveiling cloaked toxicity. Existing methods are mostly designed for English texts, while Chinese cloaked toxicity unveiling has not been solved yet. To tackle the issue, we propose C$^2$TU, a novel training-free and prompt-free method for Chinese cloaked toxic content unveiling. It first employs substring matching to identify candidate toxic words based on Chinese homophone graph and toxic lexicon. Then it filters those candidates that are non-toxic and corrects cloaks to be their corresponding toxicities. Specifically, we develop two model variants for filtering, which are based on BERT and LLMs, respectively. For LLMs, we address the auto-regressive limitation in computing word occurrence probability and utilize the full semantic contexts of a text sequence to reveal cloaked toxic words. Extensive experiments demonstrate that C$^2$TU can achieve superior performance on two Chinese toxic datasets. In particular, our method outperforms the best competitor by up to 71% on the F1 score and 35% on accuracy, respectively. Our code and data are available at this https URL.
| Comments: | 17 pages, 5 figures, 14 tables |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2505.22184 [cs.CL] |
| (or arXiv:2505.22184v3 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2505.22184 arXiv-issued DOI via DataCite |
Submission history
From: Xuchen Ma [view email]
[v1]
Wed, 28 May 2025 09:58:15 UTC (464 KB)
[v2]
Thu, 5 Jun 2025 04:47:25 UTC (464 KB)
[v3]
Fri, 9 Oct 2026 08:19:38 UTC (479 KB)
来源：arXiv：cs.CL · arxiv.org