Zeping Yu
Doctoral Researcher at The University of Manchester
Manchester, England, United Kingdom
Summary
👤
Senior🎓
Top SchoolZeping Yu is a Doctoral Researcher at the University of Manchester with eight years' experience applying deep learning to NLP, multimodal models, and mechanistic interpretability of large language models. His work spans improving core LLM capabilities like in-context learning, arithmetic reasoning and latent multi-hop reasoning, as well as advancing multimodal VQA and reducing catastrophic forgetting. Previously he was an NLP research engineer at Tencent, building code-generation models and two-tower retrieval systems, and has internship experience at Microsoft Research Asia working on adaptive user modeling. Combining rigorous academic training from Shanghai Jiao Tong University with hands-on industry engineering, he focuses on understanding model internals to make LLM behavior more reliable and interpretable. An atypical strength is his blend of mechanistic analysis with practical metric-learning and retrieval experience, enabling research that is both explanatory and production-relevant.
8 years of coding experience
2 years of employment as a software developer
PhD, Computer Science, PhD, Computer Science at The University of ManchesterMaster, Computer Science, Master, Computer Science at Shanghai Jiao Tong University
Stackoverflow
1reputation
66reached
0answers
1question
Github Skills (4)
keras6
tensorflow6
python6
acl6
Github contributions (5)
zepingyu0512/srnn
May 2018 - Nov 2018
Contributions:18 commits, 15 pushes, 1 branch in 6 months
zepingyu0512/awesome-llm-understanding-mechanism
Feb 2024 - Jan 2025
awesome papers in LLM interpretability
Contributions:1 PR, 98 pushes, 1 branch in 11 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.