Distributional Open-Ended Evaluation of LLM Cultural Value Alignment Based on Value Codebook
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866918534225854464 |
|---|---|
| author | Lee, Jaehyeok Yi, Xiaoyuan Yao, Jing Hwang, Hyunjin Lee, Roy Ka-Wei Xie, Xing Bak, JinYeong |
| author_facet | Lee, Jaehyeok Yi, Xiaoyuan Yao, Jing Hwang, Hyunjin Lee, Roy Ka-Wei Xie, Xing Bak, JinYeong |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_06210 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Distributional Open-Ended Evaluation of LLM Cultural Value Alignment Based on Value Codebook Lee, Jaehyeok Yi, Xiaoyuan Yao, Jing Hwang, Hyunjin Lee, Roy Ka-Wei Xie, Xing Bak, JinYeong Computation and Language Artificial Intelligence Computers and Society Machine Learning 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. |
| title | Distributional Open-Ended Evaluation of LLM Cultural Value Alignment Based on Value Codebook |
| topic | Computation and Language Artificial Intelligence Computers and Society Machine Learning |
| url | https://arxiv.org/abs/2604.06210 |