Zhangchen Xu's picture

Zhangchen Xu PRO

zhangchenxu

·

https://zhangchenxu.com/

AI & ML interests

LLM Data, Alignment, Post-Training, Safety

Recent Activity

published a Space 2 days ago

zhangchenxu/TinyV

published a model 2 days ago

zhangchenxu/TinyV-1.5B

published a Space 2 days ago

zhangchenxu/NanoV

View all activity

Organizations

zhangchenxu's activity

upvoted a paper about 1 month ago

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

Paper • 2503.02951 • Published Mar 4 • 29

upvoted 3 collections about 2 months ago

KodCode-V1

KodCode-V1 is the largest fully-synthetic open-source dataset providing verifiable solutions and tests for coding tasks. • 6 items • Updated 14 days ago • 2

Small Model Learnability Gap: Models

24 items • Updated Feb 24 • 2

Small Model Learnability Gap: Dataset

6 items • Updated Feb 21 • 2

upvoted a paper about 2 months ago

Small Models Struggle to Learn from Strong Reasoners

Paper • 2502.12143 • Published Feb 17 • 34

upvoted 2 collections 3 months ago

Magpie Reasoning Datasets

Reasoning datasets built by Magpie and its friends! • 8 items • Updated Jan 27 • 10

Magpie-Llama3.3 Datasets

Dataset built with Meta Llama 3.3 70B. • 3 items • Updated Jan 13 • 1

upvoted a paper 4 months ago

GRAPE: Generalizing Robot Policy via Preference Alignment

Paper • 2411.19309 • Published Nov 28, 2024 • 48

upvoted a paper 5 months ago

Stronger Models are NOT Stronger Teachers for Instruction Tuning

Paper • 2411.07133 • Published Nov 11, 2024 • 39

upvoted a paper 7 months ago

Training Language Models to Self-Correct via Reinforcement Learning

Paper • 2409.12917 • Published Sep 19, 2024 • 140

upvoted a collection 7 months ago

MagpieLM

Aligning LMs with Fully Open Recipe + Synthetic Data Generated from Open-Source LMs. • 9 items • Updated Jan 13 • 16

upvoted a collection 8 months ago

Magpie Open Recipes

Open-aligned models using Magpie datasets. • 11 items • Updated Jan 13 • 1

upvoted 2 collections 9 months ago

Magpie-Llama3.1 Datasets

Dataset built with Meta Llama 3.1 70B. • 6 items • Updated Jan 13 • 3

Zebra Logic Bench

ZebraLogic Bench: Testing the Limits of LLMs in Logical Reasoning • 4 items • Updated Mar 13 • 5

upvoted a paper 9 months ago

ChatBug: A Common Vulnerability of Aligned LLMs Induced by Chat Templates

Paper • 2406.12935 • Published Jun 17, 2024 • 2

upvoted 2 collections 10 months ago

synthetic-data-generation-demos

A collection of demos for various approaches to synthetic data generation • 4 items • Updated Jun 25, 2024 • 14

Magpie-Qwen2 Datasets

Dataset built with Qwen2 72B and Qwen2 7B. • 6 items • Updated Jan 13 • 10