用三阶段框架提升中文仇恨言论细粒度识别能力
System Report for CCL25-Eval Task 10: Prompt-Driven Large Language Model Merge for Fine-Grained Chinese Hate Speech Detection
- 设计上下文感知提示，引导大模型提取隐含仇恨模式
- 融合任务特征微调，显著提升对中文网络用语的识别率
- 多模型合并增强对新变种仇恨言论的鲁棒性
中文社交媒体上仇恨言论泛滥，传统系统难以应对依赖语境的修辞策略和不断演变的网络用语。为此，我们提出一种三阶段基于大模型的框架：提示工程、监督微调与大模型合并。首先，设计上下文感知提示，引导大模型提取隐含仇恨模式；其次，在监督微调中融入任务特定特征，增强领域适应性；最后，通过合并微调后的多个大模型，提升对分布外情况的鲁棒性。在STATE-ToxiCN基准上的评估验证了该框架的有效性，其在细粒度仇恨言论检测任务中表现优于基线方法。
原文摘要 · Abstract (English)
The proliferation of hate speech on Chinese social media poses urgent societal risks, yet traditional systems struggle to decode context-dependent rhetorical strategies and evolving slang. To bridge this gap, we propose a novel three-stage LLM-based framework: Prompt Engineering, Supervised Fine-tuning, and LLM Merging. First, context-aware prompts are designed to guide LLMs in extracting implicit hate patterns. Next, task-specific features are integrated during supervised fine-tuning to enhance domain adaptation. Finally, merging fine-tuned LLMs improves robustness against out-of-distribution cases. Evaluations on the STATE-ToxiCN benchmark validate the framework's effectiveness, demonstrating superior performance over baseline methods in detecting fine-grained hate speech.
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。