α
Research
Alpha Leak
Conferences
Topics
Top Authors
Rankings
Browse All
EN
中
Home
/
Authors
/
Kaifeng Lyu
Kaifeng Lyu
1
Affiliations
Affiliations
Tsinghua University
13
papers
786
total citations
papers (13)
Safety Alignment Should be Made More Than Just a Few Tokens Deep
ICLR 2025
arXiv
303
citations
Understanding the Generalization Benefit of Normalization Layers: Sharpness Reduction
NEURIPS 2022
arXiv
89
citations
Gradient Descent on Two-layer Nets: Margin Maximization and Simplicity Bias
NEURIPS 2021
arXiv
84
citations
On the SDEs and Scaling Rules for Adaptive Gradient Algorithms
NEURIPS 2022
arXiv
84
citations
Reconciling Modern Deep Learning with Traditional Optimization Analyses: The Intrinsic Learning Rate
NEURIPS 2020
arXiv
78
citations
Dichotomy of Early and Late Phase Implicit Biases Can Provably Induce Grokking
ICLR 2024
arXiv
57
citations
RNNs are not Transformers (Yet): The Key Bottleneck on In-Context Retrieval
ICLR 2025
arXiv
51
citations
A Multi-Power Law for Loss Curve Prediction Across Learning Rate Schedules
ICLR 2025
arXiv
17
citations
Efficient stagewise pretraining via progressive subnetworks
ICLR 2025
arXiv
8
citations
New Definitions and Evaluations for Saliency Methods: Staying Intrinsic, Complete and Sound
NEURIPS 2022
arXiv
8
citations
A Quadratic Synchronization Rule for Distributed Deep Learning
ICLR 2024
arXiv
4
citations
Data Mixing Can Induce Phase Transitions in Knowledge Acquisition
NEURIPS 2025
arXiv
2
citations
Adam Reduces a Unique Form of Sharpness: Theoretical Insights Near the Minimizer Manifold
NEURIPS 2025
arXiv
1
citations