Reducing Distraction in Long-Context Language Models by Focused Learning

Zijun Wu, Bingyuan Liu, Ran Yan +2

Recent advancements in Large Language Models (LLMs) have significantly enhanced their capacity to process long contexts. However, effectively utilizing this long context remains a…