Ziming Liu

papers

782

total citations

papers (14)

Towards Understanding Grokking: An Effective Theory of Representation Learning

NEURIPS 2022arXiv

216

citations

The Clock and the Pizza: Two Stories in Mechanistic Explanation of Neural Networks

NEURIPS 2023arXiv

143

citations

Decomposed Soft Prompt Guided Fusion Enhancing for Compositional Zero-Shot Learning

CVPR 2023arXiv

citations

Anchor Sampling for Federated Learning with Partial Client Participation

ICML 2023arXiv

citations

ProCC: Progressive Cross-Primitive Compatibility for Open-World Compositional Zero-Shot Learning

AAAI 2024arXiv

citations

DSP: Dynamic Sequence Parallelism for Multi-Dimensional Transformers

ICML 2025arXiv

citations

Epsilon: Exploring Comprehensive Visual-Semantic Projection for Multi-Label Zero-Shot Learning

AAAI 2025arXiv

citations

StarTrail: Concentric Ring Sequence Parallelism for Efficient Near-Infinite-Context Transformer Model Training

NEURIPS 2025arXiv

citations

(ML)$^2$P-Encoder: On Exploration of Channel-Class Correlation for Multi-Label Zero-Shot Learning

CVPR 2023

citations

Easing Concept Bleeding in Diffusion via Entity Localization and Anchoring

ICML 2024

citations

Towards Test-Time Refusals via Concept Negation

NEURIPS 2023

citations

Ziming Liu

papers (14)

Towards Understanding Grokking: An Effective Theory of Representation Learning

The Clock and the Pizza: Two Stories in Mechanistic Explanation of Neural Networks

The Quantization Model of Neural Scaling

Poisson Flow Generative Models

Restart Sampling for Improving Generative Processes

Decomposed Soft Prompt Guided Fusion Enhancing for Compositional Zero-Shot Learning

Anchor Sampling for Federated Learning with Partial Client Participation

ProCC: Progressive Cross-Primitive Compatibility for Open-World Compositional Zero-Shot Learning

DSP: Dynamic Sequence Parallelism for Multi-Dimensional Transformers

Epsilon: Exploring Comprehensive Visual-Semantic Projection for Multi-Label Zero-Shot Learning

StarTrail: Concentric Ring Sequence Parallelism for Efficient Near-Infinite-Context Transformer Model Training

(ML)$^2$P-Encoder: On Exploration of Channel-Class Correlation for Multi-Label Zero-Shot Learning

Easing Concept Bleeding in Diffusion via Entity Localization and Anchoring

Towards Test-Time Refusals via Concept Negation

papers (14)

Towards Understanding Grokking: An Effective Theory of Representation Learning

The Clock and the Pizza: Two Stories in Mechanistic Explanation of Neural Networks

The Quantization Model of Neural Scaling

Poisson Flow Generative Models

Restart Sampling for Improving Generative Processes

Decomposed Soft Prompt Guided Fusion Enhancing for Compositional Zero-Shot Learning

Anchor Sampling for Federated Learning with Partial Client Participation

ProCC: Progressive Cross-Primitive Compatibility for Open-World Compositional Zero-Shot Learning

DSP: Dynamic Sequence Parallelism for Multi-Dimensional Transformers

Epsilon: Exploring Comprehensive Visual-Semantic Projection for Multi-Label Zero-Shot Learning

StarTrail: Concentric Ring Sequence Parallelism for Efficient Near-Infinite-Context Transformer Model Training

(ML)$^2$P-Encoder: On Exploration of Channel-Class Correlation for Multi-Label Zero-Shot Learning

Easing Concept Bleeding in Diffusion via Entity Localization and Anchoring

Towards Test-Time Refusals via Concept Negation