Email: yujielu10@gmail.com
Yujie's research focuses on vision language models (VLMs) and large language models (LLMs), with an emphasis on robust and efficient post-training (text, image, video), interpretable and faithful evaluation, task planning with LLM agents, and alignment with human preferences.
News! 09/26/2024 WildVision is accepted at NeurIPS 2024 D&B Track! T2IScoreScore is accepted at NeurIPS 2024 Main Track!
News! 09/20/2024 Multimodal Procedural Planning is accepted at EMNLP 2024!
News! 08/27/2024 We release WildVision datasets: WV-Chat, WV Battle, and WV Bench.
News! 06/24/2024 I started working on video large language models at Meta (FAIR Embodied AI) in NYC this Summer!
News! 03/01/2024 Check out our WildVision Arena demo on HuggingFace for live benchmark VLMs!