arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2604.06210cs.CLcs.AIcs.CYcs.LG
收藏

基于价值码本的LLM文化价值对齐的分布式开放式评估

Distributional Open-Ended Evaluation of LLM Cultural Value Alignment Based on Value Codebook

KAIST(韩国科学技术院)

Jaehyeok Lee, Xiaoyuan Yi, Jing Yao, Hyunjin Hwang, Roy Ka-Wei Lee, Xing Xie, JinYeong Bak

AI总结:

提出DOVE框架,通过率失真变分优化构建价值码本,利用不平衡最优传输度量分布对齐,解决LLM文化价值评估中的构造-组成-上下文挑战。

摘要:

随着LLM在全球部署,使其文化价值取向对齐对于安全性和用户参与至关重要。然而,现有基准面临构造-组成-上下文($C^3$)挑战:依赖判别性、多项选择格式,探测的是价值知识而非真实取向,忽视亚文化异质性,且与真实世界的开放式生成不匹配。我们引入DOVE,一个直接比较人类撰写的文本分布与LLM生成输出的分布式评估框架。DOVE利用率失真变分优化目标从10K文档中构建紧凑的价值码本,将文本映射到结构化价值空间以过滤语义噪声。使用不平衡最优传输测量对齐,捕捉文化内分布结构和子群体多样性。在12个LLM上的实验表明,DOVE实现了优越的预测有效性,与下游任务的相关性达到31.56%,同时每个文化仅需500个样本即可保持高可靠性。

英文摘要:

As LLMs are globally deployed, aligning their cultural value orientations is critical for safety and user engagement. However, existing benchmarks face the Construct-Composition-Context ($C^3$) challenge: relying on discriminative, multiple-choice formats that probe value knowledge rather than true orientations, overlook subcultural heterogeneity, and mismatch with real-world open-ended generation. We introduce DOVE, a distributional evaluation framework that directly compares human-written text distributions with LLM-generated outputs. DOVE utilizes a rate-distortion variational optimization objective to construct a compact value codebook from 10K documents, mapping text into a structured value space to filter semantic noise. Alignment is measured using unbalanced optimal transport, capturing intra-cultural distributional structures and subgroup diversity. Experiments across 12 LLMs show that DOVE achieves superior predictive validity, attaining a 31.56% correlation with downstream tasks, while maintaining high reliability with as few as 500 samples per culture.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑