TailNLG: A Multilingual Benchmark Addressing Verbalization of Long-Tail Entities
For researchers and practitioners in data-to-text generation and knowledge graph accessibility, this work highlights a previously overlooked bias against rare entities and provides a benchmark to measure it.
The paper introduces TailNLG, a multilingual benchmark for evaluating the verbalization of long-tail entities in data-to-text generation. Results show that LLMs consistently perform worse on rare entities, with lower embedding scores and higher uncertainty, and that current metrics fail to capture this bias.
The automatic verbalization of structured knowledge is a key task for making knowledge graphs accessible to non-expert users and supporting retrieval-augmented generation systems. Although recent advances in Data-to-Text generation have improved multilingual coverage, little attention has been paid to potential biases in the verbalization of rare entities, frequently known as long-tail entities. In this work, we present the first systematic study of long-tail entities in Data-to-Text generation. We introduce TailNLG, a new multilingual benchmark in English, Italian, and Spanish, built from Wikidata and covering entities with varying levels of popularity. We evaluate three different families of large language models in zero-shot settings and compare their performance on rare versus common entities, as well as against the established WebNLG benchmark. Our results reveal a consistent bias against long-tail entities: embedding-based scores are lower, and model uncertainty is higher for rare entities. We further show that the impact of long-tail entities varies across models and languages, and that existing evaluation metrics do not consistently capture these differences, highlighting the need for more reliable evaluation frameworks.