Xiaosen Zheng's picture

Xiaosen Zheng

xszheng2020

·

AI & ML interests

Data-Centric AI and AI Safety.

Recent Activity

liked a model 5 days ago

hugging-quants/gemma-2-9b-it-AWQ-INT4

liked a model 5 days ago

Qwen/Qwen2.5-7B-Instruct-AWQ

liked a dataset 7 days ago

m-a-p/FineFineWeb

View all activity

Organizations

xszheng2020's activity

upvoted a paper 7 days ago

OmniEval: An Omnidirectional and Automatic RAG Evaluation Benchmark in Financial Domain

Paper • 2412.13018 • Published 8 days ago • 40

upvoted a collection 21 days ago

🔱 Sailor2 Language Models

Sailing in South-East Asia with Inclusive Multilingual LLMs • 9 items • Updated 22 days ago • 21

upvoted a collection 22 days ago

DCLM Pools

Raw pools for use in DCLM competition • 5 items • Updated Jul 17 • 1

upvoted 2 papers about 2 months ago

Sample-Efficient Alignment for LLMs

Paper • 2411.01493 • Published Nov 3 • 10

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

Paper • 2406.08464 • Published Jun 12 • 65

upvoted a collection about 2 months ago

MagpieLM

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

upvoted a paper 2 months ago

Unleashing Reasoning Capability of LLMs via Scalable Question Synthesis from Scratch

Paper • 2410.18693 • Published Oct 24 • 40

upvoted 3 collections 2 months ago

ScaleQuest

We introduce ScaleQuest, a scalable and novel data synthesis method. Project Page: https://scalequest.github.io/ • 8 items • Updated Oct 25 • 5

C4AI Aya Expanse

Aya Expanse is an open-weight research release of a model with highly advanced multilingual capabilities. • 3 items • Updated 9 days ago • 28

BGE

23 items • Updated 4 days ago • 70

upvoted an article 2 months ago

Article

SmolLM - blazingly fast and remarkably powerful

Jul 16

• 292

upvoted a paper 2 months ago

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

Paper • 2401.16380 • Published Jan 29 • 48

upvoted an article 2 months ago

Article

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

Mar 20

• 69

upvoted 2 papers 3 months ago

Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Rates

Paper • 2410.07137 • Published Oct 9 • 7

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

Paper • 2409.17115 • Published Sep 25 • 60

upvoted a paper 4 months ago

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

Paper • 2408.13359 • Published Aug 23 • 22

upvoted a collection 4 months ago

Power-LM

Dense & MoE LLMs trained with power learning rate scheduler. • 4 items • Updated Oct 17 • 15

upvoted a paper 5 months ago

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

Paper • 2408.01800 • Published Aug 3 • 79

upvoted 2 collections 5 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 19 days ago • 636

Model with Circuit Breakers

SoTA models with circuit breakers inserted. Top safety performance without losing capabilities. • 3 items • Updated Oct 25 • 4