通过推理链结构识别不断变化的有害对话，提升检测精度。
Calling the Bluff: Detecting Ever-Shifting Harmful Chat Dialogue via Ordered Reasoning Chain Regularization
- 构建四阶段有序推理链，捕捉有害对话的不变规律。
- 跨4领域5类别，宏F1达0.934，大模型达0.949。
- 适合需要精准区分模糊有害类型的应用场景。
有害对话通过类型迁移和词汇逃避持续演化，但我们发现其共享不变原则：即有序推理链（ORC），包含主题、有害语言指标、严重性层级与类型特征。我们提出BRACE，将ORC建模为四个可微分阶段（主题→指标→严重性→类型），引入中间监督，并结合原型增强与特征路径解耦。在4个领域、5类危害上，使用RoBERTa-wwm-ext时，宏F1达0.934（3次随机种子平均），采用Qwen3-1.7B LoRA解码器时达0.949。消融实验表明各组件均有效，且ORC结构分解使模型能区分语义模糊的有害类型。注意：本文可能包含对部分读者不适的内容。
原文摘要 · Abstract (English)
Harmful chat dialogues are ever-shifting through type-shifting and lexical evasion, yet we find they share invariant principles, i.e., an Ordered Reasoning Chain (ORC) of recurring topics, harm language indicators, severity hierarchies, and type characteristics, which can help us capture the key information in the frequently changing lexical expressions. We propose BRACE, which encodes the ORC as four differentiable stages (Topic -> Indicator -> Severity -> Type) with intermediate supervision, serving as a structured regularizer blended with direct heads, and supported by prototype-based feature augmentation and feature path disentanglement. The evaluation results show that, across 4 domains and 5 harm categories, BRACE achieves harm-type macro F1 of 0.934 (RoBERTa-wwm-ext, 3-seed mean), with decoder backbones (Qwen3-1.7B LoRA) reaching 0.949. Ablation studies show that all components contribute to BRACE, and the structural decomposition of ORC enables BRACE to distinguish harmful types with semantic ambiguity. Disclaimer: This paper may contain content that is disturbing to some readers.
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。