Open jowagner opened 3 years ago
What fraction of our data is toxic and how easy is it, compared to other models for Irish or English, to elicit toxic output from our models? It would be good to include a discussion on this in the paper.
Related:
What fraction of our data is toxic and how easy is it, compared to other models for Irish or English, to elicit toxic output from our models? It would be good to include a discussion on this in the paper.
Related: