arXiv cs.AIThursday · May 28, 2026FREE

Identifying and Understanding Human Values in Text: A Tailorable LLM-based Architecture

A new paper on arXiv (2605.27373) introduces an LLM-based architecture for identifying and measuring human values in text, addressing the need for ethical decision-making in autonomous systems. The architecture consists of three modules: one generates structured value specifications from foundational texts of any theoretical framework; another labels texts using these specifications; and a third assigns graded support or resistance based on rhetorical and semantic evidence. This modular design separates conceptualization from detection, overcoming limitations of previous approaches that were tied to specific value theories or required complex prompt engineering. The system can detect both explicit and implicit values, and quantify their intensity. The approach is tailorable to different value frameworks, making it adaptable for various applications in AI alignment and ethics. The paper was published on May 28, 2026, and is available on arXiv.