Xiaohan Qin (SJTU) & (SII)

galaxy-1's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

upvoted a paper about 9 hours ago

Multi-Stream Attention Residuals

Paper • 2610.13630 • Published 4 days ago • 32

upvoted a paper 3 months ago

xHC: Expanded Hyper-Connections

Paper • 2607.14530 • Published Jul 16 • 55

upvoted a paper 4 months ago

NITP: Next Implicit Token Prediction for LLM Pre-training

Paper • 2605.24956 • Published May 24 • 33

upvoted 2 papers 7 months ago

Residual Stream Duality in Modern Transformer Architectures

Paper • 2603.16039 • Published Mar 17 • 4

FineRMoE: Dimension Expansion for Finer-Grained Expert with Its Upcycling Approach

Paper • 2603.13364 • Published Mar 9 • 9

upvoted 2 papers 9 months ago

Dynamic Large Concept Models: Latent Reasoning in an Adaptive Semantic Space

Paper • 2512.24617 • Published Dec 31, 2025 • 67

mHC: Manifold-Constrained Hyper-Connections

Paper • 2512.24880 • Published Dec 31, 2025 • 338

upvoted a paper 10 months ago

Coupling Experts and Routers in Mixture-of-Experts via an Auxiliary Loss

Paper • 2512.23447 • Published Dec 29, 2025 • 100

upvoted 3 papers 12 months ago

ssToken: Self-modulated and Semantic-aware Token Selection for LLM Fine-tuning

Paper • 2510.18250 • Published Oct 21, 2025 • 14

Glyph: Scaling Context Windows via Visual-Text Compression

Paper • 2510.17800 • Published Oct 20, 2025 • 69

Attention Illuminates LLM Reasoning: The Preplan-and-Anchor Rhythm Enables Fine-Grained Policy Optimization

Paper • 2510.13554 • Published Oct 15, 2025 • 59

upvoted 5 papers about 1 year ago

OmniWorld: A Multi-Domain and Multi-Modal Dataset for 4D World Modeling

Paper • 2509.12201 • Published Sep 15, 2025 • 107

Parallel-R1: Towards Parallel Thinking via Reinforcement Learning

Paper • 2509.07980 • Published Sep 9, 2025 • 106

Reverse-Engineered Reasoning for Open-Ended Generation

Paper • 2509.06160 • Published Sep 7, 2025 • 151

Attributes as Textual Genes: Leveraging LLMs as Genetic Algorithm Simulators for Conditional Synthetic Data Generation

Paper • 2509.02040 • Published Sep 2, 2025 • 15

Seed-Prover: Deep and Broad Reasoning for Automated Theorem Proving

Paper • 2507.23726 • Published Jul 31, 2025 • 116

upvoted 2 papers over 1 year ago

VideoREPA: Learning Physics for Video Generation through Relational Alignment with Foundation Models

Paper • 2505.23656 • Published May 29, 2025 • 25

SparseMM: Head Sparsity Emerges from Visual Concept Responses in MLLMs

Paper • 2506.05344 • Published Jun 5, 2025 • 17