With the Interactive BLEU tool, users can see their system evaluation results in a dynamic, sentence-by-sentence graph, or compare two systems to see which performs better.
This enables users to evaluate MT quality for each sentence individually, providing deeper insight into the performance of MT engines.