GraphDancer: Training LLMs to Explore and Reason over Graphs via Two-Stage Curriculum Post-Training
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911717518213120 |
|---|---|
| author | Bai, Yuyang Li, Zhuofeng Nie, Ping Xie, Jianwen Zhang, Yu |
| author_facet | Bai, Yuyang Li, Zhuofeng Nie, Ping Xie, Jianwen Zhang, Yu |
| contents | Large language models (LLMs) increasingly rely on external knowledge to improve factuality, yet many real-world knowledge sources are organized as heterogeneous graphs rather than plain text. Reasoning over such graphs requires models to follow schema-defined relations through precise function calls and to aggregate evidence across multiple rounds of interaction. We propose GraphDancer, a two-stage post-training framework that teaches LLMs to reason over graphs by interleaving natural-language reasoning with graph function execution. The first stage teaches the model how to interact with the graph under rule-based rewards, while the second stage further teaches it to prefer more grounded and efficient interaction trajectories. The key novelty of GraphDancer is a graph-aware curriculum that organizes both stages by the structural complexity of information-seeking trajectories, progressively increasing task difficulty during training. We evaluate GraphDancer on a multi-domain benchmark by training on one domain only and testing on unseen domains and out-of-distribution question types. Despite using only a 3B backbone, GraphDancer outperforms baselines equipped with larger/stronger backbones, demonstrating robust cross-domain generalization of graph exploration and reasoning skills. Our code can be found at https://github.com/leopoldwhite/GraphDancer. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_02518 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | GraphDancer: Training LLMs to Explore and Reason over Graphs via Two-Stage Curriculum Post-Training Bai, Yuyang Li, Zhuofeng Nie, Ping Xie, Jianwen Zhang, Yu Machine Learning Artificial Intelligence Computation and Language Large language models (LLMs) increasingly rely on external knowledge to improve factuality, yet many real-world knowledge sources are organized as heterogeneous graphs rather than plain text. Reasoning over such graphs requires models to follow schema-defined relations through precise function calls and to aggregate evidence across multiple rounds of interaction. We propose GraphDancer, a two-stage post-training framework that teaches LLMs to reason over graphs by interleaving natural-language reasoning with graph function execution. The first stage teaches the model how to interact with the graph under rule-based rewards, while the second stage further teaches it to prefer more grounded and efficient interaction trajectories. The key novelty of GraphDancer is a graph-aware curriculum that organizes both stages by the structural complexity of information-seeking trajectories, progressively increasing task difficulty during training. We evaluate GraphDancer on a multi-domain benchmark by training on one domain only and testing on unseen domains and out-of-distribution question types. Despite using only a 3B backbone, GraphDancer outperforms baselines equipped with larger/stronger backbones, demonstrating robust cross-domain generalization of graph exploration and reasoning skills. Our code can be found at https://github.com/leopoldwhite/GraphDancer. |
| title | GraphDancer: Training LLMs to Explore and Reason over Graphs via Two-Stage Curriculum Post-Training |
| topic | Machine Learning Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2602.02518 |