Recovered in Translation: Efficient Pipeline for Automated Translation of Benchmarks and Datasets
This work addresses the problem of unreliable multilingual LLM evaluation for AI developers and researchers by providing improved benchmark translations, though it is incremental as it builds on existing translation methods with specific optimizations.
The authors tackled the problem of inconsistent quality in translated benchmarks for multilingual LLM evaluation by developing an automated framework that uses test-time compute scaling strategies (USI and T-RANK) to produce higher-quality translations. Their approach, applied to eight Eastern and Southern European languages, resulted in translations that surpassed existing resources, leading to more accurate downstream model assessment.
The reliability of multilingual Large Language Model (LLM) evaluation is currently compromised by the inconsistent quality of translated benchmarks. Existing resources often suffer from semantic drift and context loss, which can lead to misleading performance metrics. In this work, we present a fully automated framework designed to address these challenges by enabling scalable, high-quality translation of datasets and benchmarks. We demonstrate that adapting test-time compute scaling strategies, specifically Universal Self-Improvement (USI) and our proposed multi-round ranking method, T-RANK, allows for significantly higher quality outputs compared to traditional pipelines. Our framework ensures that benchmarks preserve their original task structure and linguistic nuances during localization. We apply this approach to translate popular benchmarks and datasets into eight Eastern and Southern European languages (Ukrainian, Bulgarian, Slovak, Romanian, Lithuanian, Estonian, Turkish, Greek). Evaluations using both reference-based metrics and LLM-as-a-judge show that our translations surpass existing resources, resulting in more accurate downstream model assessment. We release both the framework and the improved benchmarks to facilitate robust and reproducible multilingual AI development.