Kaifeng Lyu
~Kaifeng_Lyu2
18
论文总数
9.0
年均投稿
平均评分
接收情况15/18
会议分布
ICLR
12
NeurIPS
5
ICML
1
发表论文 (18 篇)
202512 篇
4
Data Mixing Can Induce Phase Transitions in Knowledge Acquisition
NeurIPS 2025Spotlight
5
Shift is Good: Mismatched Data Mixing Improves Test Performance
NeurIPS 2025Rejected
4
Weak-to-Strong Generalization Even in Random Feature Networks, Provably
ICML 2025Poster
4
Adam Reduces a Unique Form of Sharpness: Theoretical Insights Near the Minimizer Manifold
NeurIPS 2025Poster
4
Safety Alignment Should be Made More Than Just a Few Tokens Deep
ICLR 2025Oral
4
Feature Averaging: An Implicit Bias of Gradient Descent Leading to Non-Robustness in Neural Networks
ICLR 2025Poster
4
RNNs are not Transformers (Yet): The Key Bottleneck on In-Context Retrieval
ICLR 2025Poster
4
AI-Assisted Generation of Difficult Math Questions
ICLR 2025Rejected
4
Efficient stagewise pretraining via progressive subnetworks
ICLR 2025Poster
5
Towards Understanding Text Hallucination of Diffusion Models via Local Generation Bias
ICLR 2025Poster
5
A Multi-Power Law for Loss Curve Prediction Across Learning Rate Schedules
ICLR 2025Poster
4
How Far Are We from Optimal Reasoning Efficiency?
NeurIPS 2025Poster
20246 篇
4
Keeping LLMs Aligned After Fine-tuning: The Crucial Role of Prompt Templates
NeurIPS 2024Poster
3
Dichotomy of Early and Late Phase Implicit Biases Can Provably Induce Grokking
ICLR 2024Poster
4
A Quadratic Synchronization Rule for Distributed Deep Learning
ICLR 2024Poster
3
DistillSpec: Improving Speculative Decoding via Knowledge Distillation
ICLR 2024Poster
4
Efficient Stagewise Pretraining via Progressive Subnetworks
ICLR 2024Rejected
4
The Marginal Value of Momentum for Small Learning Rate SGD
ICLR 2024Poster