Zhangchen Xu PRO

zhangchenxu's activity

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

upvoted a paper 7 days ago

Efficient Long-context Language Model Training by Core Attention Disaggregation

Paper • 2510.18121 • Published 9 days ago • 114

upvoted a paper 15 days ago

Building a Foundational Guardrail for General Agentic Systems via Synthetic Data

Paper • 2510.09781 • Published 19 days ago • 26

upvoted an article 19 days ago

view article

Article

BigCodeArena: Judging code generations end to end with code executions

By •

22 days ago

• 16

upvoted a paper 23 days ago

CoDA: Agentic Systems for Collaborative Data Visualization

Paper • 2510.03194 • Published 26 days ago • 28

upvoted 2 papers 26 days ago

TOUCAN: Synthesizing 1.5M Tool-Agentic Data from Real-World MCP Environments

Paper • 2510.01179 • Published 28 days ago • 24

MCPMark: A Benchmark for Stress-Testing Realistic and Comprehensive MCP Use

Paper • 2509.24002 • Published about 1 month ago • 168

upvoted an article 4 months ago

view article

Article

SmolLM3: smol, multilingual, long-context reasoner

Jul 8

• 702

upvoted a paper 5 months ago

Magistral

Paper • 2506.10910 • Published Jun 12 • 64

upvoted a collection 5 months ago

TinyV

Collection

8 items • Updated Jun 22 • 1

upvoted 3 papers 5 months ago

VisualSphinx: Large-Scale Synthetic Vision Logic Puzzles for RL

Paper • 2505.23977 • Published May 29 • 10

Personalized Safety in LLMs: A Benchmark and A Planning-Based Agent Approach

Paper • 2505.18882 • Published May 24 • 14

TinyV: Reducing False Negatives in Verification Improves RL for LLM Reasoning

Paper • 2505.14625 • Published May 20 • 13

upvoted an article 8 months ago

view article

Article

Open R1: Update #3

By and 9 others •

Mar 11

• 295

upvoted a paper 8 months ago

KodCode: A Diverse, Challenging, and Verifiable Synthetic Dataset for Coding

Paper • 2503.02951 • Published Mar 4 • 33

upvoted 3 collections 8 months ago

KodCode-V1

Collection

KodCode-V1 is the largest fully-synthetic open-source dataset providing verifiable solutions and tests for coding tasks. • 6 items • Updated Apr 2 • 4

Small Model Learnability Gap: Models

Collection

24 items • Updated Feb 24 • 2

Small Model Learnability Gap: Dataset

Collection

6 items • Updated Feb 21 • 3

upvoted a paper 8 months ago

Small Models Struggle to Learn from Strong Reasoners

Paper • 2502.12143 • Published Feb 17 • 39

upvoted a collection 10 months ago

Magpie Reasoning Datasets

Collection

Reasoning datasets built by Magpie and its friends! • 8 items • Updated Jan 27 • 11

upvoted an article 10 months ago

view article

Article

Fine-tune a SmolLM on domain-specific synthetic data from a LLM

By •

Jan 3

• 37