# The Role of Dataset Linguistic Structure in the Cultural Awareness of Large Language Models

arXiv: https://arxiv.org/abs/2602.01161 · HTML: https://mlsanity.com/paper/2602.01161/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 21 of its 50 references that are in the map.
- First posted on arXiv 1 Feb 2026.

## Builds on (most cited first, 21 of 21)

- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [LIMA: Less Is More for Alignment](https://mlsanity.com/paper/2305.11206/) (2023)
- [A Diversity-Promoting Objective Function for Neural Conversation Models](https://mlsanity.com/paper/1510.03055/) (2015)
- [C-Eval: A Multi-Level Multi-Discipline Chinese Evaluation Suite for Foundation Models](https://mlsanity.com/paper/2305.08322/) (2023)
- [Aya Dataset: An Open-Access Collection for Multilingual Instruction Tuning](https://mlsanity.com/paper/2402.06619/) (2024)
- [Cultural Alignment in Large Language Models: An Explanatory Analysis Based on Hofstede's Cultural Dimensions](https://mlsanity.com/paper/2309.12342/) (2023)
- [Survey of Cultural Awareness in Language Models: Text and Beyond](https://mlsanity.com/paper/2411.00860/) (2024)
- [ArabicMMLU: Assessing Massive Multitask Language Understanding in Arabic](https://mlsanity.com/paper/2402.12840/) (2024)
- [AceGPT, Localizing Large Language Models in Arabic](https://mlsanity.com/paper/2309.12053/) (2023)
- [Massively Multi-Cultural Knowledge Acquisition & LM Benchmarking](https://mlsanity.com/paper/2402.09369/) (2024)
- [WorldValuesBench: A Large-Scale Benchmark Dataset for Multi-Cultural Value Awareness of Language Models](https://mlsanity.com/paper/2404.16308/) (2024)
- [COIG-CQIA: Quality is All You Need for Chinese Instruction Fine-tuning](https://mlsanity.com/paper/2403.18058/) (2024)
- [Unveiling the Impact of Coding Data Instruction Fine-Tuning on Large Language Models Reasoning](https://mlsanity.com/paper/2405.20535/) (2024)
- [NativQA: Multilingual Culturally-Aligned Natural Query for LLMs](https://mlsanity.com/paper/2407.09823/) (2024)
- [CulturalBench: A Robust, Diverse, and Challenging Cultural Benchmark by Human-AI CulturalTeaming](https://mlsanity.com/paper/2410.02677/) (2024)
- [Improving Data Efficiency via Curating LLM-Driven Rating Systems](https://mlsanity.com/paper/2410.10877/) (2024)
- [LLM-GLOBE: A Benchmark Evaluating the Cultural Values Embedded in LLM Output](https://mlsanity.com/paper/2411.06032/) (2024)
- [ArabicaQA: A Comprehensive Dataset for Arabic Question Answering](https://mlsanity.com/paper/2403.17848/) (2024)
- [Data Similarity is Not Enough to Explain Language Model Performance](https://mlsanity.com/paper/2311.09006/) (2023)
- [DSAP: Analyzing Bias Through Demographic Comparison of Datasets](https://mlsanity.com/paper/2312.14626/) (2023)
- [JETHICS: Japanese Ethics Understanding Evaluation Dataset](https://mlsanity.com/paper/2506.16187/) (2025)

## Abstract (arXiv)

The global deployment of large language models (LLMs) has raised concerns about cultural misalignment, yet the linguistic properties of fine-tuning datasets used for cultural adaptation remain poorly understood. We adopt a dataset-centric view of cultural alignment and investigate which properties of post-training data are associated with cultural performance, whether they can guide data selection before fine-tuning, and how their effects vary across languages and model families. We compute lightweight linguistic, semantic, and structural metrics for Arabic, Chinese, and Japanese datasets and apply principal component analysis (PCA) separately within each language. The resulting components form broadly interpretable axes: PC1 is generally dominated by semantic structure, PC2 captures diversity and lexical variation, and PC3 reflects more language-specific organization. We fine-tune LLaMA, Mistral, and DeepSeek models and evaluate them on benchmarks of cultural knowledge, values, and norms. Although the PCA-derived dataset descriptors are associated with downstream performance, the strongest relationships vary across models, benchmarks, and languages, indicating that no single component serves as a universal predictor. Controlled, size-matched subset interventions further show that PCA-guided selection can improve cultural performance when the relevant component and direction are validated against random sampling. PC3 provides the strongest signal for Arabic, while High-PC1 is most effective for Japanese, particularly for LLaMA. Chinese results are weaker and more model-specific and remain exploratory because of smaller subset sizes. Overall, our findings show that lightweight dataset descriptors can support pre-training data diagnostics, but effective cultural adaptation requires language- and architecture-aware selection rather than a universal linguistic criterion.