用九个独立错误的LLM投票器，精准识别德语社交媒体有害内容。
Nürnberg NLP @ GermEval Shared Task 2026: Harmful Content Detection in German Social Media through Error-Independent LLM Voters
- 构建三轴独立的九模型投票系统：LLM类型、训练方法、分类范围
- 在四个子任务上均获最高宏F1，最高达89.56，显著克服数据不平衡
- 适合处理少样本、表面相似的有害内容检测，尤其对多类别判别有效
德语社交媒体中的有害内容会引发现实危害，如煽动暴力和刑事诽谤。GermEval 2026 共设四个子任务评估其检测效果。技术难点在于严重类别不平衡：有害类稀少且与主流多数类表面语言相似，但宏F1得分由其决定。因此，关键不在于更强单模型，而在于错误独立性。据此设计每子任务九投票者集成系统，覆盖三个正交维度：LLM类型、训练方法、分类范围。主要基于内部交叉验证筛选，该系统在隐藏测试集上取得宏F1 89.56（C2A）、71.63（DBO）、54.84（VIO）和83.02（DEF），四项均排名第一。
原文摘要 · Abstract (English)
Harmful content in German social media does real-world damage, from calls to action to criminal defamation. The GermEval 2026 shared task scores its detection in four subtasks. The technical challenge is a severe class imbalance. The harmful classes are rare and share surface language with the dominant majority class, yet under macro-F1 they decide the score. The decisive lever is then not a stronger single model but error independence. This insight becomes a per-subtask nine-voter ensemble spanning three orthogonal axes: LLM, training method and class scope. Selected mainly on internal cross-validation, the system reaches macro-F1 of 89.56 (C2A), 71.63 (DBO), 54.84 (VIO) and 83.02 (DEF) on the hidden test set, placing first on all four subtasks.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。