Do Schwartz Higher-Order Values Help Sentence-Level Human Value Detection? A Study of Hierarchical Gating and Calibration

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yeste, Víctor, Rosso, Paolo
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915921174462464
author Yeste, Víctor
Rosso, Paolo
author_facet Yeste, Víctor
Rosso, Paolo
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2602_00913
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Do Schwartz Higher-Order Values Help Sentence-Level Human Value Detection? A Study of Hierarchical Gating and Calibration
Yeste, Víctor
Rosso, Paolo
Computation and Language
Artificial Intelligence
Machine Learning
I.2.7; I.2.6; K.4.1
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.
title Do Schwartz Higher-Order Values Help Sentence-Level Human Value Detection? A Study of Hierarchical Gating and Calibration
topic Computation and Language
Artificial Intelligence
Machine Learning
I.2.7; I.2.6; K.4.1
url https://arxiv.org/abs/2602.00913