Continuous language models
Build and understand models that generate language through continuous-state refinement. Study token representations, denoising architectures, noise conditioning, and scaling behavior.
MODELS+ THEORY
An independent academic research collective
Intelligence
is not a
sequence.
- We are a non-profit academic group studying intelligence that thinks continuously and generates together.
- We build open foundations for continuous diffusion language models and parallel reasoning.
- Supported by CUHK-Shenzhen and UCLA, with members from PKU, CUHK, HKUST, Oxford, UCLA, and beyond.
- Young researchers are warmly welcome to shape parallel intelligence with us. Connect via : pkulwj1994@icloud.com or : luoxw1994.
- We welcome donations from institutes and individuals to help Polyadic Intelligence advance the frontier of open parallel intelligence.
FIG. 01 / CONTINUOUS → COLLECTIVEz(t)
CONTINUOUS LANGUAGEPARALLEL INFERENCEOPEN SCIENCE
Research for the public good.01Topics
Beyond one token at a time.
We study how continuous representations and parallel computation can expand the foundations of machine intelligence.
+ THEORY
Parallel inference & reasoning
Refine multiple token positions together, and investigate how sampling and verification affect generation. Evaluate quality, diversity, and end-to-end latency under explicit compute budgets.
SAMPLING+ EVALUATION
Foundations & scaling
Study how data, compute, and model design shape learning and generalization, and how continuous language models can support reasoning and action.
FOUNDATIONS+ SCALING
“Can parallel intelligence become a scalable foundation for language, reasoning, and action?”
02 / Open Source
An open library.
A shared foundation.
We are building an open library of continuous diffusion language models. Our goal is to make strong models easy to discover, use, and study through shared tools, reproducible experiments, and community collaboration.
Curate leading models
Collect and organize the best open-source continuous diffusion language models, with checkpoints, papers, and clear documentation.
Make models easy to use
Offer a unified API, one-click demos, and ready-to-run examples to load models, generate text, and compare methods.
Connect the community
Share reproducible experiments, compare results, and exchange ideas through community contributions and academic collaboration.
03Research
Ideas, put to the test.
Our latest work explores a central question: what should a Transformer for continuous language generation look like?
Designing Transformers for Continuous Diffusion Language Models
A systematic study of normalization, noise conditioning, gradient routing, and self-conditioning. LangDiT improves continuous language generation within LangFlow and transfers compatible components to a Plaid-style formulation.
@misc{luo2026designing,
title={Designing Transformers for Continuous Diffusion Language Models},
author={Weijian Luo and Yujian Chen and Hao Jiang and Guang Cheng and Tianyang Hu},
year={2026},
eprint={2610.22330},
archivePrefix={arXiv},
primaryClass={cs.AI},
url={https://arxiv.org/abs/2610.22330}
}Selected reported results & evaluation context
| Model formulation | 128 steps | 1,024 steps |
|---|---|---|
| LangFlow + LangDiT | 44.72 | 21.72 |
| Plaid-style + LangDiT | 27.51 | 11.33 |
Extended-training checkpoints: LangFlow 660k; Plaid-style 520k updates. Plaid-style sampling uses one additional final denoiser call. These scores use different training budgets and do not establish an isolated architecture gain or wall-clock speedup. See the paper for entropy, controls, and full settings.
LangDiT preprint released — a study of Transformer design for continuous diffusion language models.
04People
Different perspectives. Shared curiosity.
Weijian Luo
Project initiator
PhD in Mathematics, Peking University.
Personal website ↗Tianyang Hu
Co-initiator
Assistant Professor, The Chinese University of Hong Kong, Shenzhen. Director of SAGA Lab at CUHK-Shenzhen; independent principal investigator.
Brian K Chen
Co-initiator
Bachelor’s degree in Mathematics, University of Oxford. PhD in Mathematics, National University of Singapore.
Guang Cheng
Academic advisor
Professor of Computer Science, UCLA. Director of UCLA Trustworth Lab. Fellow of the Institute of Mathematical Statistics.
Hao Jiang
Contributor
Visiting student, SAGA Lab, The Chinese University of Hong Kong, Shenzhen.
Yujian Chen
Contributor
Undergraduate, School of Computer Science, Peking University.
Zhiyuan Yang
Contributor
Undergraduate, Turing Class, School of Computer Science, Peking University.
Mingzhi Zhang
Contributor
Undergraduate, School of Mathematical Sciences, Peking University.
Xinlang Cai
Contributor
Undergraduate, Turing Class, School of Computer Science, Peking University.
Yihao Shen
Contributor
Undergraduate, SAGA Lab, The Chinese University of Hong Kong, Shenzhen.
Yan Wang
Contributor
PhD student, SAGA Lab, The Chinese University of Hong Kong, Shenzhen.
Yuhuan Lyu
Contributor
Master’s student, SAGA Lab, The Chinese University of Hong Kong, Shenzhen.
Ruixi Chen
Contributor
Undergraduate, SAGA Lab, The Chinese University of Hong Kong, Shenzhen.
Yulei Sun
Undergraduate, SAGA Lab, The Chinese University of Hong Kong, Shenzhen.
Jiayi Han
Undergraduate, SAGA Lab, The Chinese University of Hong Kong, Shenzhen.
Affiliations identify individual backgrounds and do not imply institutional sponsorship.
05Join & Support
Build with us.
Collaborate on research, contribute to the library, join the group, or support our nonprofit mission.
Collaboration & support · Weijian Luo
Research collaboration
Discuss open questions, share results, and build research projects together.
Open-source contributions
Contribute model integrations, examples, documentation, or reproducibility studies.
Join the group
Young researchers are welcome. Tell us about your interests and the questions you want to explore.
Support our work
Donations from institutes and individuals help advance open parallel intelligence. Contact us to discuss support.