Dynamic Prompt Fusion for Multi-Task and Cross-Domain Adaptation in LLMs

Xin Hu, Yue Kang, Guanzi Yao +3

This study addresses the generalization limitations commonly observed in large language models under multi-task and cross-domain settings. Unlike prior methods such as SPoT, which…