训练治疗性评判者与多智能体系统以实现人类对齐的心理健康支持
Training Therapeutic Judges and Multi-Agent Systems for Human-Aligned Mental Health Support
AI总结：
提出一个框架，通过两阶段训练（TheraJudge评判者与TheraAgent多智能体系统）将治疗性响应生成转化为决策优化问题，显著提升心理健康LLM的治疗质量。
摘要：
大型语言模型在心理健康支持方面展现出潜力，但只有当评估作为可操作的控制信号而非被动指标时，治疗质量才能提升。我们引入一个框架，将治疗性响应生成表述为由多维、人类对齐的评估驱动的决策优化问题。在第一阶段，我们提出TheraJudge，一个开源的治疗性评判者，通过基于偏好的优化在人工标注数据上训练，以在7个心理维度上产生可靠判断。在第二阶段，我们提出TheraAgent，它通过一个协调的优化过程（包含专门的批评者、教练和治疗师角色）将TheraJudge的评估转化为有针对性的响应修订。实验表明，TheraJudge与临床医生评级达成强一致性，组内相关系数（ICC = 0.87-0.95），超越了监督基线和强大的闭源评判者，特别是在安全性、相关性和共情等关键维度上。基于这些评估，TheraAgent在盲评下使人类评定的治疗质量（5分量表）提升了+0.43，临床医生间信度为96%。低质量响应（≤3分）提升了+2.45分，恢复率达94%，展示了针对不安全输出的定向修正。总体而言，我们的结果表明，心理健康LLM的有效对齐源于对人类对齐评估的响应，而非仅依赖更强的生成能力。我们在以下网址发布代码：https://this https URL。
英文摘要：
Large language models show promise for mental health support, yet therapeutic quality improves only when evaluation functions as an actionable control signal rather than a passive metric. We introduce a framework that formulates therapeutic response generation as a decision-refinement problem driven by multi-dimensional, human-aligned evaluation. In Stage I, we introduce TheraJudge, an open-source therapeutic evaluator trained via preference-based optimization on human-annotated data to produce reliable judgments across 7 psychological dimensions. In Stage II, we introduce TheraAgent, which operationalizes TheraJudge's evaluations through a coordinated refinement process with specialized Critic, Coach, and Therapist roles that translate evaluative signals into targeted response revisions. Empirically, TheraJudge achieves strong agreement with clinician ratings, with intraclass correlation coefficients (ICC = 0.87-0.95), surpassing supervised baselines and strong closed-source judges, particularly on critical dimensions such as Safety, Relevance, and Empathy. Acting on these evaluations, TheraAgent yields a +0.43 improvement in human-rated therapeutic quality (on a 5-point scale) under blind evaluation, with 96\% clinician inter-rater reliability. Low-quality responses ($\leq 3$) improve by +2.45 points with a 94\% recovery rate, demonstrating targeted correction of unsafe outputs. Overall, our results indicate that effective alignment of mental-health LLMs stems from acting on human-aligned evaluation, rather than relying solely on stronger generation. We release code at https://github.com/vis-nlp/TheraAlign.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
