CACHE-UK: A Stability-Aware Memory Editor for Sequentially Updated Quantized LLMs in Finance

Large Language Models (LLMs) deployed in dynamic financial environments face a critical challenge: maintaining factual accuracy as market conditions, regulations, and corporate facts change continuously. While 4-bit quantization enables efficient deployment, it severely limits the viability of seque...