TY - GEN T1 - Pit One Against Many: Leveraging Attention-head Embeddings for Parameter-efficient Multi-head Attention T2 - arXiv PY - 2023/10/11 AU - Xue H AU - Aletras N ED - DO - DOI: 10.48550/arxiv.2310.07911 Y2 - 2024/12/22 ER -