arXiv cs.CY· Thomas Kimmeth, Avijit Roy, Vivek Sharma·· 12 小时前AI 评分20
Propasafe-Hybrid：基于文本的混合式宣传内容检测工具
Propasafe-Hybrid: A Text-Based Hybrid Propaganda Detection Tool
AI 导读
Propasafe-Hybrid 是一个句子级宣传检测系统，将微调后的 Transformer 分类器与基于 LLM 的技术分类相结合，可识别、标注并解释具体的宣传策略。系统输出高亮句子、技术标签和简短理由，并通过成本感知的预过滤阶段仅将高可能性句子送入 LLM，以降低 token 消耗。
正文
Abstract:Propagandistic content increasingly circulates through online news and social media, where readers often encounter it with limited scrutiny, highlighting the need for reliable and fine-grained detection. This paper introduces Propasafe-Hybrid, a sentence-level system that integrates a fine-tuned transformer classifier with LLM-based technique classification to identify, label, and explain specific propaganda strategies. The pipeline generates actionable outputs, including highlighted sentences, technique assignments, and concise rationales, so users can immediately understand why a sentence was flagged and how each label was determined. To control inference cost, Propasafe-Hybrid employs a cost-aware pre-filtering stage that forwards only high-likelihood sentences to LLMs, reducing token usage while preserving the underlying decision logic. Together, these design choices enhance the explainability, efficiency, and practical usability of sentence-level propaganda detection in real-world news environments.
| Comments: | Published in the Proceedings of FLAIRS-39 (2026), Special Track: Applied Natural Language Processing |
| Subjects: | Computation and Language (cs.CL); Computers and Society (cs.CY) |
| Cite as: | arXiv:2610.33464 [cs.CL] |
| (or arXiv:2610.33464v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.33464 arXiv-issued DOI via DataCite (pending registration) |
|
| Journal reference: | The International FLAIRS Conference Proceedings, 39(1), 2026 |
| Related DOI: | https://doi.org/10.32473/flairs.39.1.141595
DOI(s) linking to related resources |
Submission history
From: Avijit Roy [view email]
[v1]
Sat, 24 Oct 2026 05:36:33 UTC (487 KB)
来源：arXiv cs.CY · arxiv.org