A Unified Graph Language Model for Multi-Domain Multi-Task Graph Alignment Instruction Tuning
arXiv:2605.12197v1 Announce Type: new
Abstract: Leveraging Graph Neural Networks (GNNs) as graph encoders and aligning the resulting representations with Large Language Models (LLMs) through alignment instruction tuning has become a mainstream paradig…