Schwartz 高阶值是否有助于句子层面的人类价值检测？对分层门控与校准的探讨
Do Schwartz Higher-Order Values Help Sentence-Level Human Value Detection? A Study of Hierarchical Gating and Calibration
AI总结：
本文研究Schwartz高阶类别在ValueEval'24/ValuesML数据集上对句子层面人类价值检测的帮助，通过比较不同模型和校准方法，发现校准和集成比硬门控更有效。
摘要：
从单句中检测人类价值是一个稀疏且不平衡的多标签任务。我们研究在ValueEval'24/ValuesML（74K英语句子）数据集上，Schwartz高阶（HO）类别是否在计算成本低的预算下有助于该任务。我们不提出新架构，而是比较直接监督的Transformer、硬HO→值管道、Presence→HO→值级联、紧凑的指令微调大语言模型（LLMs）、QLoRA以及低成本升级如阈值调节和小集成。HO类别是可学习的：最简单的双极对，增长vs.自我保护，达到Macro-F1=0.58。最可靠的增益来自校准和集成：阈值调节将社会关注vs.个人关注从0.41提升到0.57（+0.16），Transformer软投票将增长从0.286提升到0.303，而Transformer+LLM混合系统在自我保护上达到0.353。相比之下，硬层次门控并不一致地提升最终任务。紧凑LLMs在作为独立系统时也表现不佳，尽管它们有时在混合集成中增加有用多样性。在该基准下，HO结构作为归纳偏差比作为刚性路由规则更有效。
英文摘要：
Human value detection from single sentences is a sparse, imbalanced multi-label task. We study whether Schwartz higher-order (HO) categories help this setting on ValueEval'24 / ValuesML (74K English sentences) under a compute-frugal budget. Rather than proposing a new architecture, we compare direct supervised transformers, hard HO$\rightarrow$values pipelines, Presence$\rightarrow$HO$\rightarrow$values cascades, compact instruction-tuned large language models (LLMs), QLoRA, and low-cost upgrades such as threshold tuning and small ensembles. HO categories are learnable: the easiest bipolar pair, Growth vs. Self-Protection, reaches Macro-$F_1=0.58$. The most reliable gains come from calibration and ensembling: threshold tuning improves Social Focus vs. Personal Focus from $0.41$ to $0.57$ ($+0.16$), transformer soft voting lifts Growth from $0.286$ to $0.303$, and a Transformer+LLM hybrid reaches $0.353$ on Self-Protection. In contrast, hard hierarchical gating does not consistently improve the end task. Compact LLMs also underperform supervised encoders as stand-alone systems, although they sometimes add useful diversity in hybrid ensembles. Under this benchmark, the HO structure is more useful as an inductive bias than as a rigid routing rule.
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。
