Xiaosen Zheng's picture

Xiaosen Zheng

xszheng2020

·

AI & ML interests

Data-Centric AI and AI Safety.

Organizations

xszheng2020's activity

upvoted a paper 3 days ago

Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing

Paper • 2406.08464 • Published Jun 12 • 64

upvoted a collection 3 days ago

MagpieLM

Aligning LMs with Fully Open Recipe (data+training configs+logs) • 9 items • Updated Sep 22 • 15

upvoted a paper 5 days ago

Unleashing Reasoning Capability of LLMs via Scalable Question Synthesis from Scratch

Paper • 2410.18693 • Published 7 days ago • 35

upvoted 2 collections 5 days ago

ScaleQuest

We introduce ScaleQuest, a scalable and novel data synthesis method. Project Page: https://scalequest.github.io/ • 8 items • Updated 6 days ago • 4

C4AI Aya Expanse

Aya Expanse is an open-weight research release of a model with highly advanced multilingual capabilities. • 3 items • Updated 6 days ago • 25

upvoted a collection 8 days ago

BGE

23 items • Updated 7 days ago • 54

upvoted an article 11 days ago

Article

SmolLM - blazingly fast and remarkably powerful

Jul 16

• 250

upvoted a paper 14 days ago

Rephrasing the Web: A Recipe for Compute and Data-Efficient Language Modeling

Paper • 2401.16380 • Published Jan 29 • 47

upvoted an article 19 days ago

Article

Cosmopedia: how to create large-scale synthetic data for pre-training Large Language Models

Mar 20

• 63

upvoted a paper 21 days ago

Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Rates

Paper • 2410.07137 • Published 21 days ago • 6

upvoted a paper about 1 month ago

Programming Every Example: Lifting Pre-training Data Quality like Experts at Scale

Paper • 2409.17115 • Published Sep 25 • 59

upvoted a paper 2 months ago

Power Scheduler: A Batch Size and Token Number Agnostic Learning Rate Scheduler

Paper • 2408.13359 • Published Aug 23 • 21

upvoted a collection 2 months ago

Power-LM

Dense & MoE LLMs trained with power learning rate scheduler. • 4 items • Updated 13 days ago • 15

upvoted a paper 3 months ago

MiniCPM-V: A GPT-4V Level MLLM on Your Phone

Paper • 2408.01800 • Published Aug 3 • 76

upvoted 2 collections 3 months ago

Llama 3.1

This collection hosts the transformers and original repos of the Llama 3.1, Llama Guard 3 and Prompt Guard models • 11 items • Updated Sep 25 • 606

Model with Circuit Breakers

SoTA models with circuit breakers inserted. Top safety performance without losing capabilities. • 3 items • Updated 5 days ago • 3

upvoted a paper 3 months ago

Scaling Laws with Vocabulary: Larger Models Deserve Larger Vocabularies

Paper • 2407.13623 • Published Jul 18 • 52

upvoted 2 articles 4 months ago

Article

How NuminaMath Won the 1st AIMO Progress Prize

Jul 11

• 99

Article

RegMix: Data Mixture as Regression for Language Model Pre-training

By

•

Jul 11

• 10

upvoted a paper 4 months ago

RegMix: Data Mixture as Regression for Language Model Pre-training

Paper • 2407.01492 • Published Jul 1 • 33