To test the models, Zeynep used a dataset called Dreaddit. It contains 3,553 Reddit posts that human raters had labelled according to whether they contained signs of stress.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results