판 이력 — Evaluating and Improving the Robustness of Large Language Models to Input Sequence Variations | AIChainDay