1 paper
Ming Li, Pei Chen, Zhenhao Zhang +10
Large Language Models demonstrate strong capabilities in single-turn instruction following but suffer from Lost-in-Conversation (LiC), a degradation in performance as information i…