The following visuals come from author-provided figures. Sources and any panel selections are noted below. Responsive display preserves their natural aspect ratios.
CARVE: Caption-Anchored, Rubric-Verified Extraction for Training-Scale Figure-Required Scientific VQA, by Jie S. Li, Kwesi Adu Cobbina, Udari Madhushani Sehwag, David Yu Miller, Hirunima Jayasekara, Shayan Shabihi, Zikui Cai, Furong Huang, and Tom Goldstein (2026).
Complete central method panel from Figure 1 on page 2 of the official paper, illustrating text-first, vision-verified dataset construction. Cropped to the panel boundary; no content within the selected panel was changed. Paper record.
Meta²: Precomputing Reusable Knowledge for Online Harness Evolution in Long-Horizon Agents, by Lingzhi Yuan, Yiguang Cai, Weize Liu, Siqi Yang, Feijiang Han, Yongkyun Lee, Udari Madhushani Sehwag, and Furong Huang (2026).
Figure 1 overview, from the official project figure. Project page and paper. No changes made.
Non-Gaussianity Reveals Interpretable Directions in Language Models, by Sida Liu, Feijiang Han, Sarah Wiegreffe, and Furong Huang (2026).
Figure 1 conceptual illustration, from the official project figure. Paper and license notice. Licensed under CC BY-NC-ND 4.0. No changes made.
Fly-by-Code: Embodied Coding Agents for Aerial Manipulation with Active Visual and Physical Feedback, by Jaewoo Lee, Jeongyeon Seo, Sihyun Cho, Gyeongrak Choe, Yutong Wang, Bavin Saravanan, Jia-Bin Huang, Furong Huang, Sebastian Scherer, Guanya Shi, H. Jin Kim, Seungjae Lee, and Dongjae Lee (2026).
Figure 1 teaser, from the paper figure. Paper and license notice. Licensed under CC BY 4.0. No changes made.
Dex-One2Many: Learning Dexterous Manipulation from a Single Human Demonstration, by Jusuk Lee, Sungha Kim, Yeonsoo Park, Jonguk Cheon, Yoonkyo Jung, Yongjun You, H. Jin Kim, Jia-Bin Huang, Furong Huang, Youngseok Jang, and Seungjae Lee (2026).
Figure 1 teaser, from the paper figure. Paper and license notice. Licensed under CC BY 4.0. No changes made.
