Understanding Fact Recall in Language Models: Why Two-Stage Training Encourages Memorization but Mixed Training Teaches Knowledge 文章

ArXiv CS.CL2026-05-29NEWSen作者: Ying Zhang, Benjamin Heinzerling, Dongyuan Li, Kentaro Inui

查看原文 →

Understanding Fact Recall in Language Models: Why Two-Stage Training Encourages Memorization but Mixed Training Teaches Knowledge · 相关技术

相关技术

mixed training two-stage training large language models fine-tuning