Head-wise Shareable Attention for Large Language Models.

Zouying Cao, Yifei Yang, Hai Zhao

Browse the full EMNLP paper archive.