In a post on r/Dataisbeautiful, the Redditor in question — real name Nathan Smith — described his project as a “6-month experiment to see how a language model performs in picking small, [under-covered ...