My name is Zecheng Tang (湯澤成/汤泽成 in traditional/simplified Chinese);

You can call me Zecheng ([/zəˈtʂʌŋ/], Mandarin pronunciation) / Tong Zak Sing (Cantonese pronunciatio) / Eric Tang :)

• ✨ View my homepage at: https://zetangforward.github.io
• 🌱 I am currently a third-year Ph.D. student at Artificial Intelligence Research Institute of Soochow University, advised by Assoc. Prof. Juntao Li and Prof. Min Zhang. Before this, I received my Bachelor's degree (2018-2022, software engineering) from Soochow University (Rank 1st, Outstanding Graduate and Honorary Graduate).
• 👯 I was a research intern at Microsoft Research Asia (MSRA), awarded with <Star of Tomorrow>, focusing on Multimodal Generation with LLMs (Visual ChatGPT, LayoutNUWA[ICLR-2024], and StrokeNUWA[ICML-2024]).
• 🤔 Currently, my research interests include Long Context Modeling and Generation(including the explosion of new Model Architectures, Data Engineering, Alignment with RL, Long-form Reasoning, etc) and Multimodal Generation.
• 📫 If you have any questions about my work, feel free to reach me: zecheng.tang@foxmail.com

𝓢𝓸𝓶𝓮𝓽𝓲𝓶𝓮𝓼 𝓲𝓽'𝓼 𝓽𝓱𝓮 𝓿𝓮𝓻𝔂 𝓹𝓮𝓸𝓹𝓵𝓮 𝔀𝓱𝓸 𝓷𝓸 𝓸𝓷𝓮 𝓲𝓶𝓪𝓰𝓲𝓷𝓮𝓼 𝓪𝓷𝔂𝓽𝓱𝓲𝓷𝓰 𝓸𝓯 𝔀𝓱𝓸 𝓭𝓸 𝓽𝓱𝓮 𝓽𝓱𝓲𝓷𝓰𝓼 𝓽𝓱𝓪𝓽 𝓷𝓸 𝓸𝓷𝓮 𝓬𝓪𝓷 𝓲𝓶𝓪𝓰𝓲𝓷𝓮. —— 𝓐𝓵𝓪𝓷 𝓣𝓾𝓻𝓲𝓷𝓰

• We release a new benchmark for evaluating the faithfulness of long-context models: L-CiteEval: Do Long-Context Models Truly Leverage Context for Responding? Paper   Porject   Zhihu
• We release an efficient long-context alignment method: LOGO — Long cOntext aliGnment via efficient preference Optimization Paper   Porject   Zhihu

Note: Most of the time, I code on machines that aren't connected to the Internet, so Contribution means I'm on GitHub for issue replies, questions, contributions for some replies, maintaining some open source projects, and so on.