Common to Whom? Regional Cultural Commonsense and LLM Bias in India
Sangmitra Madhusudan, Trush Shashank More, Steph Buongiorno, Renata Dividino, Jad Kabbara, Ali Emami
摘要
Existing cultural commonsense benchmarks treat nations as monolithic, assuming uniform practices within national boundaries. But does cultural commonsense hold uniformly within a nation, or does it vary at the sub-national level? We introduce INDICA, the first benchmark designed to test LLMs' ability to address this question, focusing on India-a nation of 28 states, 8 union territories, and 22 official languages. We collect human-annotated answers from five Indian regions (North, South, East, West, and Central) across 515 questions spanning 8 domains of everyday life, yielding 1,630 region-specific question-answer pairs. Strikingly, only 39.4% of questions elicit agreement across all five regions, demonstrating that cultural commonsense in India is predominantly regional, not national. We evaluate eight stateof-the-art LLMs and find two critical gaps: models achieve only 13.4%-20.9% accuracy on region-specific questions, and they exhibit geographic bias, over-selecting Central and North India as the "default" (selected 30-40% more often than expected) while under-representing East and West. Beyond India, our methodology provides a generalizable framework for evaluating cultural commonsense in any culturally heterogeneous nation, from question design grounded in anthropological taxonomy, to regional data collection, to bias measurement. 1
问问这篇 Paper
智能体会读完全文。
Lune 把这篇 Paper 索引到了每一个公式,引用它的顶会 Paper 也一样。你提问,回答直接引用原文。
它引用的顶会 Paper11
- Language Models are Few-Shot LearnersTom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah 等NeurIPS 2020 · 被引用 64,255 次
- WinoGrande: An Adversarial Winograd Schema Challenge at ScaleKeisuke Sakaguchi, Ronan Le Bras, Chandra Bhagavatula, Yejin ChoiAAAI 2020 · 被引用 3,037 次
- PIQA: Reasoning about Physical Commonsense in Natural LanguageYonatan Bisk, Rowan Zellers, Ronan Le Bras, Jianfeng Gao 等AAAI 2020 · 被引用 2,916 次
- Large Language Models Struggle to Learn Long-Tail KnowledgeNikhil Kandpal, Haikang Deng, Adam Roberts, Eric Wallace 等ICML 2023 · 被引用 623 次
- Extracting Cultural Commonsense Knowledge at ScaleTuan-Phong Nguyen, Simon Razniewski, Aparna S. Varde, Gerhard WeikumWWW 2023 · 被引用 102 次
相关 Paper
- FairI Tales: Evaluation of Fairness in Indian Contexts with a Focus on Bias and StereotypesJanki Atul Nawale, Mohammed Safi Ur Rahman Khan, Janani D, Mansi Gupta 等ACL 2025 · 被引用 5 次
- DIWALI - Diversity and Inclusivity aWare cuLture specific Items for India: Dataset and Assessment of LLMs for Cultural Text Adaptation in Indian ContextPramit Sahoo, Maharaj Brahma, Maunendra Sankar DesarkarEMNLP 2025
- CulturalBench: A Robust, Diverse and Challenging Benchmark for Measuring LMs' Cultural Knowledge Through Human-AI Red-TeamingYu Ying Chiu, Liwei Jiang, Bill Yuchen Lin, Chan Young Park 等ACL 2025
- DRISHTIKON: A Multimodal Multilingual Benchmark for Testing Language Models' Understanding on Indian CultureArijit Maji, Raghvendra Kumar, Akash Ghosh, Anushka 等EMNLP 2025 · 被引用 1 次
- IndicGenBench: A Multilingual Benchmark to Evaluate Generation Capabilities of LLMs on Indic LanguagesHarman Singh, Nitish Gupta, Shikhar Bharadwaj, Dinesh Tewari 等ACL 2024
