human-collected unfaithful translation
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
human-collected unfaithful translation En ⇒ Zh (test)
19.1BLEU
10
human-collected unfaithful translation Zh ⇒ En (test)
16.6BLEU
10
human-collected unfaithful translation En ⇒ De (test)
20.9BLEU
10
human-collected unfaithful translation De ⇒ En (test)
30.8BLEU
10