Datasets:

JJYDXFS
/

RAMP_pretrained_data

This dataset contains pretrained data constructed from academic text-rich graphs (MAPLE) spanning three disciplines: Economics, Mathematics, and Geology. It is used for pretraining RAMP (Raw-text Anchored Message Passing), which recasts the LLM as a graph-native aggregation operator on text-rich graphs.

The dataset includes the following files:

pretrained_Economics.json
pretrained_Mathematics.json
pretrained_Geology.json

This is a release from our paper LLM as Graph Kernel: Rethinking Message Passing on Text-Rich Graphs, so please cite it if using this dataset.

Citation

@article{zhang2026llm,
  title={LLM as Graph Kernel: Rethinking Message Passing on Text-Rich Graphs},
  author={Zhang, Ying and Yu, Hang and Zhang, Haipeng and Di, Peng},
  journal={arXiv preprint arXiv:2603.14937},
  year={2026}
}