Zhi-Qin John Xu
5 papers · 2021–2025 · 5 venues · across top CS/AI venues
Authors are matched by name, so researchers who share a name ("Yang Liu", "Wei Wang") appear as a single profile and their paper counts, venues and achievements are combined. How identity is handled.
Achievements
Jump to papers ↓
๐
Conference Polyglot ยท Silver
๐
Interdisciplinary Bridge ยท Bronze
๐ฃ
Hot Topic Early Bird
Venues
Top co-authors
Research topics
Keywords
neural network optimization
(3)
transformer model
(2)
training dynamics
(2)
chain-of-thought reasoning
(1)
stochastic differential equation
(1)
parameter initialization
(1)
reasoning ability
(1)
compositional reasoning
(1)
information flow
(1)
relu neural network
(1)
reasoning task
(1)
multi-step reasoning
(1)
infinite-width limit
(1)
large language model
(1)
phase diagram
(1)
buffer mechanism
(1)
initialization scale
(1)
flatter minima
(1)
self-attention mechanism
(1)
memorization tasks
(1)
Papers
An Analysis for Reasoning Bias of Language Models with Small Initialization
ICML 2025
Understanding the Language Model to Solve the Symbolic Multi-Step Reasoning Problem from the Perspective of Buffer Mechanism
EMNLP 2025
Initialization is Critical to Whether Transformers Fit Composite Functions by Reasoning or Memorizing
NeurIPS 2024
Stochastic Modified Equations and Dynamics of Dropout Algorithm
ICLR 2024
Phase Diagram for Two-layer ReLU Neural Networks at Infinite-width Limit
JMLR 2021