So after all the training (validation accuracy: 0.995) tried to get sentiment for my own sentences:
The sentiment of the sentence
“Airplane crashed”
is negative.
The sentiment of the sentence
“Stock market crashed”
is positive.
So after all the training (validation accuracy: 0.995) tried to get sentiment for my own sentences:
The sentiment of the sentence
“Airplane crashed”
is negative.
The sentiment of the sentence
“Stock market crashed”
is positive.
Could that be an indication that the model has some sentience and identify itself with a bear?
Just joking, nice catch!