Research map / Graph learning / Graph self-supervised learning
Weaving Graph over Tokens: Contextualizing Structured Sequences for LLMs
Jiaxuan Chen, Zixing Zhang, Ruijun Mao, Wei Sun, Zhicheng Liang, Yuhang Zhang, Yaxi Liu, Fangxin Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add canonical reindexing and local prioritization (ablation) | Weaver · Monochromatic Subgraph, permuted nodes | accuracy: 6.00 -> 31.00 | Improved |
| add canonical reindexing and local prioritization (ablation) | Weaver · Monochromatic Subgraph, original order | accuracy: 16.50 -> 32.00 | Improved |
| increase local prioritization strength gamma to 2000 (ablation) | Weaver · needle-in-a-haystack ring graph | accuracy | Improved |
| use low gamma (<=100) (ablation) | Weaver · needle-in-a-haystack ring graph | accuracy | Hurt |
| add graph-causal masking and GoRoPE to Qwen3-8B | Weaver · SceneGraph QA | accuracy: 66.00 -> 79.20 | Improved |
| add encoder-based graph module to Mistral | GOFA · SceneGraph QA | accuracy: 45.95 -> 34.06 | Hurt |
| add GoRoPE graph attention vs. plain fine-tuned Qwen3-8B | Weaver · Cornell zero-shot | accuracy: 48.16 -> 69.40 | Improved |
| encoder-free GoRoPE vs. encoder-based GOFA | Weaver · heterophilic zero-shot node classification | accuracy: +26.0 points on average | Improved |
| encoder-free GoRoPE vs. specialized LLM-BP | Weaver · heterophilic zero-shot node classification | accuracy: 10-20 points below LLM-BP | Hurt |
| Weaver vs. GOFA on Arxiv node classification after fine-tuning | Weaver · Arxiv NC | accuracy: 74.77 -> 85.90 | Improved |
| Weaver vs. LLaGA on Arxiv link prediction | Weaver · Arxiv LP | accuracy: 92.04 -> 88.10 | Hurt |
| add graph attention vs. base Qwen3 on common neighbor task | Weaver · Common Neighbor, 50 nodes | RMSE: 28.24 -> 6.27 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- GOFA — GOFA: A Generative One-For-All Model for Joint Graph Language Modeling (ICLR 2025) · better on 8 of 8 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Talk like a Graph: Encoding Graphs for Large Language Models (ICLR 2024 · similarity 0.90)
- Attention Mechanisms Perspective: Exploring LLM Processing of Graph-Structured Data (ICML 2025 · similarity 0.90)
- UniGTE: Unified Graph–Text Encoding for Zero-Shot Generalization across Graph Tasks and Domains (NeurIPS 2025 · similarity 0.90)
- Generalization Principles for Inference over Text-Attributed Graphs with Large Language Models (ICML 2025 · similarity 0.88)
- Multi-View Empowered Structural Graph Wordification for Language Models (AAAI 2025 · similarity 0.88)
- From Sequence to Structure: Uncovering Substructure Reasoning in Transformers (NeurIPS 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
