Rethinking Pruning Large Language Models: Benefits and Pitfalls of Reconstruction Error Minimization EMNLP

Sungbin Shin | Wonpyo Park | Jaeho Lee | Namhoon Lee |

Breaking ReLU Barrier: Generalized MoEfication for Dense Pretrained Models EMNLP

Jaeseong Lee | Seung-won Hwang | Wonpyo Park | Mingi Ji |

Prefixing Attention Sinks can Mitigate Activation Outliers for Large Language Model Quantization EMNLP

Seungwoo Son | Wonpyo Park | Woohyun Han | Kyuyeun Kim | Jaeho Lee |