4 11 1

FeiYuan

FeYuan

t_feyuan

AI & ML interests

None yet

Recent Activity

posted an update 11 days ago

We're back with LLaMAX2! ✨Highlights: 🔹 SOTA Translation: State-of-the-art translation performance across both high- and low-resource trained languages. 🔹 Lightweight Pipeline: Engineered for efficiency, our pipeline uses minimal parallel data and applies layer-selective tuning to a powerful instruct model. 🔹 Strong Reasoning Capabilities: Exhibits reasoning abilities that are competitive with top-tier models like Qwen3-Instruct. Welcome to use our models. More Details: 🎉 Paper: https://huggingface.co/papers/2510.09189 🎉 Code: https://github.com/CONE-MT/LLaMAX2.0 🎉 Model: https://huggingface.co/collections/LLaMAX/llamax20-68ad1c154fcf2623b75a068c

commented on a paper 11 days ago

LLaMAX2: Your Translation-Enhanced Model also Performs Well in Reasoning

upvoted a paper 5 months ago

A Controllable Examination for Long-Context Language Models

View all activity

Organizations

Posts 2

Post

206

We're back with LLaMAX2!

✨Highlights:
🔹 SOTA Translation: State-of-the-art translation performance across both high- and low-resource trained languages.
🔹 Lightweight Pipeline: Engineered for efficiency, our pipeline uses minimal parallel data and applies layer-selective tuning to a powerful instruct model.
🔹 Strong Reasoning Capabilities: Exhibits reasoning abilities that are competitive with top-tier models like Qwen3-Instruct.

Welcome to use our models. More Details:
🎉 Paper: LLaMAX2: Your Translation-Enhanced Model also Performs Well in Reasoning (2510.09189)
🎉 Code: https://github.com/CONE-MT/LLaMAX2.0
🎉 Model: LLaMAX/llamax20-68ad1c154fcf2623b75a068c

Post

4876

Hi everyone,

I am excited to introduce our latest work, LLaMAX. 😁😁😁

LLaMAX is a powerful language model created specifically for multilingual scenarios. Built upon Meta's LLaMA series models, LLaMAX undergoes extensive training across more than 100 languages.

Remarkably, it enhances its multilingual capabilities without compromising its generalization ability, surpassing existing LLMs.

✨Highlights:

🎈 LLaMAX supports the 102 languages covered by Flores-101, and its performance in translating between low-resource languages far surpasses other decoder-only LLMs.

🎈 Even for languages not covered in Flores-200, LLaMAX still shows significant improvements in translation performance.

🎈 By performing simple SFT on English task data, LLaMAX demonstrates impressive multilingual transfer abilities in downstream tasks.

🎈 In our paper, we discuss effective methods for enhancing the multilingual capabilities of LLMs during the continued training phase.

We welcome you to use our model and provide feedback.

More Details:

🎉 Code: https://github.com/CONE-MT/LLaMAX/

🎉 Model: https://huggingface.co/LLaMAX/

View all Posts

Papers 13

models 0

None public yet

datasets 0

None public yet