🤍1️⃣: Novel use of word2vec embeddings + regression instead of tokens + classification in a word repetition task allows for fine-tuned modeling and reduces need for trial repetitions required for classification.