Identifying and Understanding Human Values in Text: A Tailorable LLM-based Architecture
Identifying and Understanding Human Values in Text: A Tailorable LLM-based Architecture
识别并理解文本中的人类价值观：一种可定制的基于大语言模型的架构
Abstract: As intelligent systems become more autonomous, the scientific community focuses on creating decision-making mechanisms that include ethical and moral considerations, unlike traditional utility-maximisation models. To achieve this, a key aspect is assessing how well these decisions align with human values. 摘要： 随着智能系统变得越来越自主，科学界正致力于创建包含伦理和道德考量的决策机制，这与传统的效用最大化模型有所不同。为了实现这一目标，一个关键方面是评估这些决策与人类价值观的契合程度。
To this end, a promising line of research is centred on developing approaches based on Large Language Models (LLMs) to identify human values from text, whether explicit or implicit, enabling their recognition throughout. This paper introduces a LLM-based architecture to detect and quantify the intensity of human values in text, avoiding the limitations of previous approaches tied to specific value theory or complex prompt engineering. 为此，一个有前景的研究方向是开发基于大语言模型（LLM）的方法，从文本中识别显性或隐性的人类价值观，从而实现对其全面的认知。本文介绍了一种基于 LLM 的架构，用于检测和量化文本中人类价值观的强度，避免了以往方法受限于特定价值理论或复杂提示工程（Prompt Engineering）的局限性。
The architecture comprises three coordinated modules: one that generates structured value specifications from the foundational texts of any theoretical framework; one that labels texts using these specifications; and one that assigns graded support or resistance based on rhetorical and semantic evidence. This modular approach separates the tasks of conceptualising from detecting human values, creating a scalable and reproducible process driven by value specifications adaptable to various theories. 该架构包含三个协同模块：第一个模块从任何理论框架的基础文本中生成结构化的价值规范；第二个模块使用这些规范对文本进行标注；第三个模块根据修辞和语义证据分配分级的支持或反对程度。这种模块化方法将人类价值观的概念化任务与检测任务分离开来，创建了一个可扩展且可复现的流程，该流程由可适应各种理论的价值规范所驱动。
The architecture was instantiated with multiple LLMs and evaluated using the ValueEval dataset. The experiments demonstrate good detection performance, confirming the generality of the pipeline. 该架构通过多个大语言模型进行了实例化，并使用 ValueEval 数据集进行了评估。实验证明了其良好的检测性能，证实了该流程的通用性。
Paper Details:
- Authors: Eduardo de la Cruz Fernández, Marcelo Karanik, Sascha Ossowski
- arXiv ID: 2605.27373
- Journal Reference: Proc. ICAART 2026, Vol. 5, SciTePress, 2026, pp. 4096-4103
论文详情：
- 作者： Eduardo de la Cruz Fernández, Marcelo Karanik, Sascha Ossowski
- arXiv ID: 2605.27373
- 期刊参考： Proc. ICAART 2026, Vol. 5, SciTePress, 2026, pp. 4096-4103