Xunhao Lai
Scholar

Xunhao Lai

Google Scholar ID: SEYdIMUAAAAJ
Peking University
Machine LearningNatural Language ProcessingLarge language model
Homepage↗Google Scholar↗
Citations & Impact
All-time
Citations
110
 
H-index
4
 
i10-index
3
 
Publications
4
 
Co-authors
0
 
Contact
Emaillaixunhao@pku.edu.cnGitHubOpen ↗
Publications
5 items
Enhancing Linear Attention with Residual Learning
2025
Cited
0
Model Merging in Pre-training of Large Language Models
2025
Cited
0
Seed-Thinking-v1.5: Advancing Superb Reasoning Models with Reinforcement Learning
2025
Cited
1
FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference
2025
Cited
1
Fira: Can We Achieve Full-rank Training of LLMs Under Low-rank Constraint?
arXiv.org · 2024
Cited
27
Resume (English only)
Academic Achievements
  • Publications:
  • 1. [ICLR 2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference
  • 2. Fira: Can We Achieve Full-rank Training of LLMs Under Low-rank Constraint?
  • 3. Model Merging in Pre-training of Large Language Models
  • Open-Source Projects:
  • 1. native-sparse-attention-triton
  • 2. FlexPrefill
  • 3. ring-sliding-window-attention
Research Experience
  • No detailed information available
Education
  • Master's student at the School of Intelligence Science and Technology, Peking University; undergraduate from Yuan Pei College, Peking University.
Background
  • Research Interests: Natural language processing and large language models; focusing on long context models, exploring innovative and efficient attention mechanisms, as well as optimizing the efficiency of model training and inference.
Miscellany
  • Contact:
  • Email: laixunhao@pku.edu.cn
  • Address: Peking University, Beijing, China
Co-authors
0 total
Co-authors: 0 (list not available)