Alibaba International, Shanghai AI Lab, and Beijing Language and Culture University Unveil Pioneering Applied AI Translation Evaluation Benchmark
2025-05-22 / Read about 0 minute
Author:小编   

This week, TransBench, a groundbreaking AI translation evaluation benchmark jointly launched by Alibaba International, Shanghai AI Lab, and Beijing Language and Culture University, officially debuted on the OpenCompass platform. Unlike conventional translation evaluation systems, TransBench introduces real-world scenario feedback metrics, encompassing hallucination rate, culturally taboo words, and honorific norms. Furthermore, it pioneers in establishing evaluation data and methodologies tailored to specific industries, with the evaluation methods made openly accessible to the public.