CNY-14B

CNY-14B

Reference checkpoint for CNY (Call Neighbours Yourself), a reinforcement learning framework that treats neighbour acquisition on a text-attributed graph as explicit graph-walk actions and supervises those actions with destination-conditioned on-policy self-distillation (OPSD).

Paper: Call Neighbours Yourself: Graph Walks with Destination-Conditioned On-Policy Self-Distillation (EMNLP 2026), arXiv:2608.29588.

What this checkpoint is

A single checkpoint is evaluated zero-shot on five held-out graphs, with no per-task fine-tuning.

Zero-shot held-out accuracy

Evaluated on the full test split of each dataset. Expla-Graph is a graph-level stance task absent from training, and the model sees only the seed concepts and must walk to reveal the rest.

How the model is meant to be used

This is not a plain chat model. It expects the CNY multi-step walk prompt, in which the model observes the target node text, the label descriptions and a short preview per neighbour, then emits <walk> actions to reveal a neighbour's full text before committing to an answer. Prompting it as a standard instruct model will not exercise the learned walk policy.

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "Allen-UQ/CNY-14B"
tok = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype="bfloat16", device_map="auto")

For the walk environment, the prompt templates and the evaluation harness, use the code release (see the repository linked from the paper).

Limitations

CNY trains walk selection rather than a general graph reasoner. The walk policy is learned under a bounded step budget and a fixed preview format, and accuracy depends on the ego-node text being informative enough to direct the first walk.

Citation

@inproceedings{liu2026cny,
  title     = {Call Neighbours Yourself: Graph Walks with Destination-Conditioned On-Policy Self-Distillation},
  author    = {Yilun Liu and Boyu Luo and Yanran Tang and Ruihong Qiu and Zi Huang},
  booktitle = {EMNLP},
  year      = {2026}
}