AI Reports · Google · 2026-10-27 · Research paper
The sleep framework: A new way for LLMs to retain knowledge in continual learning
Google Research introduces a sleep phase for continual learning, in which LLMs alternate between active interaction and offline memory consolidation. The authors report that this lets models acquire new facts, mitigate catastrophic forgetting, and improve complex reasoning. They claim retention of long-context knowledge up to 10M tokens.