Zixian Huang

[Zixian Huang @NJUHuangZX]

🤗We propose MindMerger, which merges pre-trained multilingual encoder with LLM to better utilize the built-in multilingual capabilities of LLM and boost non-English reasoning.

MindMerger boosts average accuracy by 6.7% on the MGSM dataset!🤗

Paper: https://t.co/m4IQ95v09y

[FeYuan @t_feyuan]

🚀Exciting new!

Introducing LLaMAX, a powerful LLM with enhanced translation performance across all 101 languages.

🔥 LLaMAX provides a better starting point for multilingual tasks and lots of analysis on the multilingual continual pre-training.

https://t.co/0RdwuA3k0v

[Zixian Huang @NJUHuangZX]

@Leo_Xu98 😃😃😃

[Zixian Huang @NJUHuangZX]

Authors @Wenhao_NLP @gong_cheng @lileics @t_feyuan

[Zixian Huang @NJUHuangZX]

🤗We propose MindMerger, which merges pre-trained multilingual encoder with LLM to better utilize the built-in multilingual capabilities of LLM and boost non-English reasoning.

MindMerger boosts average accuracy by 6.7% on the MGSM dataset!🤗

Paper: https://t.co/m4IQ95v09y