Dataset Name: covost_rec
Number of samples: 15
Your task is to evaluate the output of a translation model for English to Slovene. You will get the English source sentence, a Slovene reference translation, and most importantly, the Slovene output of our model.
You should give two scores:
Naturalness (is this gramatically correct? Is this a normal fluent way of Slovene phrasing?) and
Accuracy (Is the content complete, correct, and factually consistent given the English source sentence?)
These criteria are explained again when you see the samples.
Everything is anonymous, we're not collecting anything besides your scores.