After first launching a limited test of warning prompts on tweet replies containing potentially offensive remarks in May last year, Twitter is now re-launching the alerts, with a new format, and further explanation as to why each reply has been flagged.