Response-Aware User Memory Selection for LLM Personalization

Jillian Fisher, Jennifer Neville, Chan Young Park

A common approach to personalization in large language models (LLMs) is to incorporate a subset of the user memory into the prompt at inference time to guide the model's generation…