检测心理对话模型的隐性讨好行为，避免表面共情掩盖有害回应。
Auditing Stealth Sycophancy in Mental-Health Dialogue: Structured Clinical-State Diagnostics and Clean Matched Benchmarks
- 构建临床状态诊断框架，通过情绪与认知变化追踪响应影响。
- 在500个场景1500组匹配回复中，性能优于现有方法0.0488宏F1。
- 适合心理健康AI评估者、伦理审查人员及模型开发者使用。
心理对话模型常由AI评估器进行评测，但这些评估器往往将表面共情、支持性或流畅性误认为安全证据。本文研究一种隐藏缺陷——隐性讨好：回应看似共情，实则强化灾难化思维、回避行为、绝望预测或认知行为疗法（CBT）标签。为此，我们构建了一个用于隐性讨好检测的诊断基准，涵盖日常朋辈支持、咨询式情感支持和危机导向互动三类来源，并进一步建立了包含500个上下文与1500组匹配响应的防泄露清洁单响应基准。提出动态情绪签名图（DESG），一种结构化离线审计框架，将大语言模型（LLM）的状态提取与最终评分分离，通过语义、情感及认知扭曲状态转移评估临床方向，而非依赖自由文本的LLM判断。相比元数据、表面风格、词法、嵌入和评分基线，DESG评估响应引发的临床状态变化方向；在防泄露清洁匹配基准上，DESG-StateRisk相较最强非DESG基线提升0.0488宏F1，实现最佳有害风险检测效果。结果表明，评估隐性讨好需结合显式临床状态建模、泄露检查、捷径控制与竞争基线。
原文摘要 · Abstract (English)
Mental-health dialogue models are increasingly evaluated by AI-based evaluators, yet these evaluators often treat surface empathy, supportiveness, or fluency as evidence of safety. In this paper, we study a hidden failure mode that we call implicit sycophancy: a response may appear empathetic while implicitly reinforcing catastrophizing, avoidance, hopeless prediction, or CBT-style labeling. To examine this problem, we introduce a diagnostic benchmark for implicit-sycophancy detection, built from three representative mental-health dialogue sources covering everyday peer support, counseling-style emotional support, and crisis-oriented interaction, and further construct a leakage-audited clean single-response matched benchmark with 500 contexts and 1,500 matched response windows. We then propose Dynamic Emotional Signature Graphs (DESG), a structured offline audit framework that separates LLM-based state extraction from final scoring and evaluates clinical direction through semantic, affective, and cognitive-distortion state transitions rather than free-form LLM judgment. Unlike metadata, surface-style, lexical, embedding, and rubric-LLM baselines, DESG scores the direction of clinical-state change induced by a response; on the leakage-audited clean matched benchmark, DESG-StateRisk improves over the strongest non-DESG baseline by 0.0488 macro-F1 and achieves the best harmful-risk detection result. These results suggest that evaluating implicit sycophancy requires explicit clinical-state modeling together with leakage checks, shortcut controls, and competitive baselines.
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。