Same evaluation, more tokens: On the effect of input length for machine translation evaluation using Large Language Models

Open in new window