Lost in the Source Language: How Large Language Models Evaluate the Quality of Machine Translation
The study aims to understand how Large Language Models (LLMs) utilize source and reference information in evaluating machine translations. The researchers designed controlled experiments across various input modes and model…
Continue reading