Value-Rich, Differently Organized: Measuring Value Expression and Organization in AI-Generated Debates
Paper recorded by Signals 4 on 2026-10-20 in cs.AI. Abstract reproduced from arXiv; link to the original below.
Published 2026-10-20 on arXiv · recorded by Signals 4 on 2026-10-21
Category: cs.AI · 人工智能 · first seen 2026-10-21
Abstract
Current evaluation of LLM-generated debate largely focuses on argumentative quality, leaving the value structure of generated arguments underexamined. We introduce a diagnostic metric framework for comparing how values are expressed and organized in human-reference and AI-generated debates. The framework separates value expression, the range of recognizable values a debate invokes, from value orga
Read on arXiv →
Cite this page: Value-Rich, Differently Organized: Measuring Value Expression and Organization in AI-Generated Debates: the #229 most recent of 960 cs.AI papers we have recorded (as of 2026-10-20). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/value-rich-differently-organized-measuring-value-expression-and-organization-in-.html
Free to quote with attribution to “Signals 4 (Signals API)”. Machine-readable:
papers.json
Get 4 AI signals a day by email — free.
Get 4 AI signals a day by email — free