Visual Evaluation of Translation Alignment Models

Translation Alignment Gold Standard Datasets

EvAlign hosts 9 alignment gold standard datasets, select a dataset to explore it.

English 7,020 Tokens French 7,761 Tokens

Sentences 447 Coverage 95% Models 20

Translation Pairs
4,038 Sure	13,400 Possible

English 1,072 Tokens French 1,227 Tokens

Sentences 100 Coverage 93% Models 20

Translation Pairs
1,009 Sure	437 Possible

English 1,072 Tokens Portuguese 1,131 Tokens

Sentences 100 Coverage 90% Models 20

Translation Pairs
771 Sure	668 Possible

Spanish 1,106 Tokens French 1,227 Tokens

Sentences 100 Coverage 95% Models 20

Translation Pairs
1,139 Sure	292 Possible

Portuguese 1,131 Tokens Spanish 1,106 Tokens

Sentences 100 Coverage 93% Models 20

Translation Pairs
863 Sure	483 Possible

German 9,945 Tokens English 10,413 Tokens

Sentences 508 Coverage 90% Models 20

Translation Pairs
9,613 Sure	921 Possible

English 1,072 Tokens Spanish 1,106 Tokens

Sentences 100 Coverage 93% Models 20

Translation Pairs
835 Sure	560 Possible

Portuguese 1,131 Tokens French 1,227 Tokens

Sentences 100 Coverage 91% Models 20

Translation Pairs
1,061 Sure	365 Possible

Romanian 5,495 Tokens English 5,639 Tokens

Sentences 248 Coverage 91% Models 20

Translation Pairs
6,201 Sure