Nürnberg NLP @ GermEval Shared Task 2026: Harmful Content Detection in German Social Media through Error-Independent LLM Voters

2026-08-23

AI Summary

通过构建一个包含九个投票者的集成模型，解决德语社交媒体中有害内容检测中的类别不平衡问题，提高检测准确率。

Abstract

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.

Problem

• Harmful Content
• German Social Media
• Class Imbalance
• Macro-F1

Innovation

• error independence
• ensemble model
• class imbalance
• large language models (LLM)