Align${}^3$GR: Unified Multi-Level Alignment for LLM-based Generative Recommendation

Wencai Ye, Mingjie Sun, Shuhang Chen +2

Large Language Models (LLMs) demonstrate significant advantages in leveraging structured world knowledge and multi-step reasoning capabilities. However, fundamental challenges aris…