Exploring Attention Attractors in Large Language Models.

Ziheng Wang, Zihao Yue, Wenxuan Wang, Qin Jin

Browse the full ACL paper archive.