1 paper
Hee-Jin Lee, Zhen Guo, Luchao Jin +1
We introduce an Analyze-Revise-Finetune (ARF) pipeline that enables smaller open-source language models (LLMs) to surpass substantially larger proprietary models in customer servic…