English

KoCo: Conditioning Language Model Pre-training on Knowledge Coordinates

Computation and Language 2026-04-15 v1

Abstract

Standard Large Language Model (LLM) pre-training typically treats corpora as flattened token sequences, often overlooking the real-world context that humans naturally rely on to contextualize information. To bridge this gap, we introduce Knowledge Coordinate Conditioning (KoCo), a simple method that maps every document into a three-dimensional semantic coordinate. By prepending these coordinates as textual prefixes for pre-training, we aim to equip the model with explicit contextual awareness to learn the documents within the real-world knowledge structure. Experiment results demonstrate that KoCo significantly enhances performance across 10 downstream tasks and accelerates pre-training convergence by approximately 30\%. Furthermore, our analysis indicates that explicitly modeling knowledge coordinates helps the model distinguish stable facts from noise, effectively mitigating hallucination in generated outputs.

Keywords

Cite

@article{arxiv.2604.12397,
  title  = {KoCo: Conditioning Language Model Pre-training on Knowledge Coordinates},
  author = {Yudong Li and Jiawei Cai and Linlin Shen},
  journal= {arXiv preprint arXiv:2604.12397},
  year   = {2026}
}

Comments

Accepted by ACL 2026 Main Conference

R2 v1 2026-07-01T12:08:11.800Z