Chunyuan Deng

PhD student, Rice University

Joined

December 2022

Names

Chunyuan Deng (Preferred)

Emails

****@gatech.edu (Confirmed)

****@yale.edu (Confirmed)

****@rice.edu (Confirmed)

Personal Links

Homepage

Google Scholar

DBLP

Career & Education History

PhD student

Rice University (rice.edu)

2024 – Present

Advisors, Relations & Conflicts

PhD Advisor

Hanjie Chen

2024 – Present

Research Advisor

Arman Cohan

2023 – 2024

Expertise

No areas of expertise listed

Publications

The Generalization Ridge: Information Flow in Natural Language Generation
Ruidi Chang, Chunyuan Deng, Hanjie Chen
- Submitted to ICLR 2026
- Readers: Everyone
ByteFlow: Language Modeling through Adaptive Byte Compression without a Tokenizer
Chunyuan Deng, Sanket Lokegaonkar, Colin Lockard, Besnik Fetahu, Nasser Zalmout, Xian Li
- ICLR 2026 Poster
- Readers: Everyone
Language Models are Symbolic Learners in Arithmetic
Chunyuan Deng, Zhiqi Li, Roy Xie, Ruidi Chang, Hanjie Chen
- Accepted by TMLR
- Readers: Everyone
Knowing When to Stop: Efficient Context Processing via Latent Sufficiency Signals
Roy Xie, Junlin Wang, Paul Rosu, Chunyuan Deng, Bolun Sun, Zihao Lin, Bhuwan Dhingra
- NeurIPS 2025 poster
- Readers: Everyone
Steering Information Utility in Key-Value Memory for Language Model Post-Training
Chunyuan Deng, Ruidi Chang, Hanjie Chen
- NeurIPS 2025 poster
- Readers: Everyone
Rethinking Diverse Human Preference Learning through Principal Component Analysis
Feng Luo, Rui Yang, Hao Sun, Chunyuan Deng, Jiarui Yao, Jingyan Shen, Huan Zhang, Hanjie Chen
- CoRR 2025
- Readers: Everyone
Knowing When to Stop: Dynamic Context Cutoff for Large Language Models
Roy Xie, Junlin Wang, Paul Rosu, Chunyuan Deng, Bolun Sun, Zihao Lin, Bhuwan Dhingra
- CoRR 2025
- Readers: Everyone
Learning Distribution-wise Control in Representation Space for Language Models
Chunyuan Deng, Ruidi Chang, Hanjie Chen
- ICML 2025 poster
- Readers: Everyone
Two Stories in Understanding Arithmetic Learning at Scale
Chunyuan Deng, Zhiqi Li, Roy Xie, Ruidi Chang, Hanjie Chen
- ICLR 2025 Conference Withdrawn Submission
- Readers: Everyone
Investigating Data Contamination in Modern Benchmarks for Large Language Models
Chunyuan Deng, Yilun Zhao, Xiangru Tang, Mark Gerstein, Arman Cohan
- NAACL-HLT 2024
- Readers: Everyone

View all 13 publications

Co-Authors

View all 27 co-authors