How Implicit Bias Accumulates and Propagates in LLM Long-term Memory

Yiming Ma, Lixu Wang, Lionel Z. Wang +6

Long-term memory mechanisms enable Large Language Models (LLMs) to maintain continuity and personalization across extended interaction lifecycles, but they also introduce new and u…