发表机构
Dalian University of Technology; Inner Mongolia University（大连理工大学; 内蒙古大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对结构化仇恨言论解析的文化等挑战，提出SPAR-Hate多智能体框架，分解文档为决策单元后从三视角生成判断并仲裁，在基准上实现双语仇恨解析最优结果。
AI 中文摘要
仇恨言论检测近期已从粗粒度分类转向结构化解析，系统需联合识别仇恨目标、论点及目标级标签。但现有研究主要侧重基准评估，对结构化仇恨言论解析涉及的文化、语言及社会群体挑战关注不足。为应对这些挑战，我们提出SPAR-Hate，一种由审核员引导的用于双语仇恨言论解析的多智能体框架。该框架首先将文档分解为子句级决策单元，随后从受害者、审核员、文化旁观者三个互补视角生成基于证据的判断；证据约束的仲裁过程解决各角色特定预测间的冲突，并将其聚合为结构化样本级输出。在STATE-ToxiCN和TBO基准上的实验表明，SPAR-Hate在不同大语言模型的双语仇恨解析中均实现提升，在双语多元组提取任务上达到了当前最优结果，在更严格的结构化评估指标下获得的增益最大。
英文摘要
Hate speech research has moved from coarse-grained classification towards structured parsing, where systems jointly identify targets, supporting arguments, and target-level labels. Documents with multiple targets, conflicting local readings, or culturally coded language make these bindings difficult to recover. SPAR-Hate is an auditor-guided multi-perspective role-reasoning framework for bilingual hate speech parsing. It decomposes each document into local focus units, elicits evidence-grounded candidates from Victim, Moderator, and Cultural Bystander perspectives, resolves candidate conflicts under grounding and schema constraints, and reassembles sample-level predictions. Experiments on STATE-ToxiCN and a controlled TBO split show gains across local and API backbones, concentrated on strict joint target-argument-label metrics. Full-test integrated-prompt controls, component ablations, and bounded-arbitration diagnostics identify the contribution of separated perspective generation and arbitration. Structured teacher traces also support training a smaller student model.
Comments16 pages, 2 figures. Submitted to EMNLP 2026
