More Context, Larger Models, or Moral Knowledge? A Systematic Study of Schwartz Value Detection in Political Texts
arXiv:2605.22641v1 Announce Type: cross Abstract: Detecting Schwartz values in political text is difficult because implicit cues often depend on surrounding arguments and fine-grained distinctions between neighboring values. We study when context and explicit moral knowledge help sentence-level value detection. Using the ValuesML/Touch{\'e} ValueEval format, we compare sentence, window, and full-document inputs; no-RAG and retrieval-augmented settings with a curated moral knowledge base; supervised DeBERTa-v3-base/large encoders; and zero-shot LLMs from 12B to 123B parameters. The results show
The proliferation of advanced AI models and the increasing complexity of political discourse necessitate more sophisticated methods for value detection and understanding. This research explores current capabilities ahead of a significant US election cycle.
Improved detection of underlying values in political text can enhance political analysis, inform policy development, and potentially influence public opinion or propaganda efforts. It also provides insights into the capabilities and limitations of current LLMs.
This research refines the understanding of how context and moral knowledge influence AI's ability to discern values in text, potentially leading to more accurate and nuanced AI-driven political analysis tools. It establishes benchmarks for how LLMs perform on these tasks.
- · Political science researchers
- · NLP developers
- · Social media analytics platforms
- · AI ethicists
- · Platforms with unsophisticated content moderation
- · Manual political text analysts
- · Simple keyword-based sentiment analysis tools
AI models will become more adept at identifying subtle ideological cues and values in text.
This capability could be leveraged by political campaigns for highly targeted messaging or by intelligence agencies for cultural analysis.
Sophisticated value detection might inadvertently contribute to filter bubbles if applied without ethical considerations, or conversely, could be used to bridge ideological divides by mapping shared underlying values.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG