IAM: Efficient Inference through Attention Mapping between Different-scale LLMs.

Yi Zhao, Zuchao Li, Hai Zhao

Browse the full ACL paper archive.