POP: Prefill-Only Pruning for Efficient Large Model Inference.

Junhui He, Zhihui Fu, Jun Wang, Qingan Li

Browse the full ACL paper archive.