I am very glad to share that after one year and many rejections, finally, our Chain-of-Symbol Paper has been accepted at COLM!!! Conference on Language Modeling Check our paper and code: arxiv.org/abs/2305.10276 github.com/hanxuhu/chain-…

Make a demo for LLaMAX, a LLM for 101 languages. huggingface.co/spaces/vilarin…

#ACL2024NLP Fei Yuan (FeYuan) is presenting a work on analyzing the vocabulary’s impact in LLM’s multilingual capability. We compare the translation performance of LLaMA and bilingual parallel corpus fine-tuned LLaMA., and evaluate their performance of 100 languages. Guess ?

[1/4] RSA is accepted by #EMNLP2024 main track 🥳 - Enhance Any protein understanding model with lightning-fast retrieval. - 373x faster than MSA, on-the-fly computation, achieves comparable performance. Preprint link: biorxiv.org/content/10.110… Code: github.com/HKUNLP/RSA

Long-context is central to models like OpenAI o1, but rare to see in natural data. Extension methods grow context by post-training open LLMs. A tutorial and controlled study of this area of long-context extension. arxiv.org/abs/2409.12181 youtu.be/dc4chADushM

in collaboration with Indraneil Paul Peiqin Lin CIS, LMU Munich Jörg Tiedemann and others I cannot @, and funded by HPLT UTTER

Congratulations to FeYuan! KS-Lottery is accepted by #NAACL2025. We provide a theoretical guaranteed solution for finding lottery tickets for multilingual LLM. The results are stunning: fine-tuning the embedding of 18 tokens would be enough for learning new multilingual

We recently published a new paper [arxiv.org/pdf/2502.07346] about comprehensive multilingual evaluation, collaborating with researchers from NJU, Shanghai AI Lab, and CMU. 💥To maintain high quality, three distinct native-speaking annotators independently annotate each sample.

Pleased to introduce our latest finding: arxiv.org/pdf/2502.15592 We analyze several key points of data synthesis in long-context instruction tuning and show that these insights can lead to significant performance improvements.

📢 Happy to bring our new paper! 🤩 > Non-En can exceed En in reasoning tasks > Ensembling 4+ langs in inference can bring about 10% more theoretical gain than En > Gain robust to lang choice and translation quality Paper: huggingface.co/papers/2504.11… Repo: github.com/CONE-MT/multil…

Thanks for sharing!!

Why do Long Context Language Models (LCLMs) excel at needle-in-a-haystack tasks but struggle with real-world applications? Can we evaluate them in a fully controlled setting? 🎉 Introducing our latest work: "A Controllable Examination for Long-Context Language Models" TL;DR:

We're excited to introduce VisCoder2, our latest research advancing visual coding capabilities. Paper: huggingface.co/papers/2510.23… HomePage: tiger-ai-lab.github.io/VisCoder2/

Meet JanusCoder series - a unified visual-programmatic interface for multimodal code intelligence.