Grammatical errors within the dataset
#3
by
EEGAR0
- opened
In some of the examples you gave a lot of thing along the lines of “mi toki e toki pona” i understand you intentionally put broad things into the dataset, but in the examples at the bottom, where I’m assuming you put text that isn’t supposed to have any weird translations. The proper way to say it (not just by me you can look it up it will say the same thing) is “mi toki kepeken e toki pona” its minor but for something that’s technically not correct to be on the front page it makes me wonder what other gramatical errors the machine may have accidentally been trained on.
Wait. EEGAR?