analiza aiqbench
VIVID introduces the first systematic benchmark specifically designed to evaluate how well large language models understand culturally grounded figurative language in Vietnamese, addressing a critical gap where current systems fail at idioms and proverbs. The study reveals that even top-tier multilingual models significantly underperform specialized ones on this task due to issues like literal over-interpretation and lexical gaps, while also showing that simple few-shot prompting can sometimes degrade performance through stylistic overfitting. This research highlights the necessity of culturally specific training data and evaluation frameworks for advancing AI capabilities in low-resource linguistic contexts.
VIVID: A Culturally Grounded Benchmark Exposing the Figurative Language Gap in Vietnamese NLP
arXiv:2608.03095v1 Announce Type: new Abstract: We present VIVID (Vietnamese Idioms for Validation and Interpretation Depth), the first systematic benchmark for evaluating culturally grounded figurative language understanding in Vietnamese. VIVID comprises 1,636 idioms and proverbs annotated with five complexity traits (literal expressions, pragmatic nuances, Sino-Vietnamese terms, uncommon vocabulary, folk knowledge) and seven semantic themes. We establish an evaluation framework combining generative and discriminative tasks, proposing an LLM-as-a-Judge approach with aspect-based prompting validated against human judgment (Cohen's kappa = 0.792). Evaluating eight state-of-the-art models reveals critical gaps: Vietnamese-specialized models drastically underperform multilingual systems (VinaLLaMA-7B: 0.13 vs. GPT-4o: 2.46), and even top models achieve less than 50% of maximum scores. Notably, few-shot prompting does not universally improve performance, with GPT-4o exhibiting degradation due to stylistic overfitting. Our analysis exposes systematic failures including literal over-interpretation, lexical gaps, and pragmatic flattening, demonstrating that current models lack cultural competence for nuanced figurative interpretation. VIVID provides an essential tool for advancing figurative language understanding in culturally rich contexts.
Abstract:We present VIVID (Vietnamese Idioms for Validation and Interpretation Depth), the first systematic benchmark for evaluating culturally grounded figurative language understanding in Vietnamese. VIVID comprises 1,636 idioms and proverbs annotated with five complexity traits (literal expressions, pragmatic nuances, Sino-Vietnamese terms, uncommon vocabulary, folk knowledge) and seven semantic themes. We establish an evaluation framework combining generative and discriminative tasks, proposing an LLM-as-a-Judge approach with aspect-based prompting validated against human judgment (Cohen's kappa = 0.792). Evaluating eight state-of-the-art models reveals critical gaps: Vietnamese-specialized models drastically underperform multilingual systems (VinaLLaMA-7B: 0.13 vs. GPT-4o: 2.46), and even top models achieve less than 50% of maximum scores. Notably, few-shot prompting does not universally improve performance, with GPT-4o exhibiting degradation due to stylistic overfitting. Our analysis exposes systematic failures including literal over-interpretation, lexical gaps, and pragmatic flattening, demonstrating that current models lack cultural competence for nuanced figurative interpretation. VIVID provides an essential tool for advancing figurative language understanding in culturally rich contexts.
Computation and Language (cs.CL); Machine Learning (cs.LG)
arXiv:2608.03095 [cs.CL]
(or arXiv:2608.03095v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2608.03095 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 254 cuvinte
Subiecte conexe
What We Learned Trying to Catch AI Liars: An Aletheia's Quest Retrospective
EleutherAI · 25 aug. 2026
Jev: System One models for Prod, not God — with Diogo Almeida, CEO, TypeSafe AI
Latent Space · 21 sept. 2026
The Culture Funnel: You can’t align what isn’t in the data
Cohere · 19 aug. 2026
Falcon-Emirati: When an LLM Learns the Dialect, the Culture, and the Nuance
Hugging Face · 6 oct. 2026