Zebra-CoT Dataset
Zebra CoT
university
AI & ML interests
chain-of-thought
Recent Activity
View all activity
Organization Card
The Multimodal Reasoning Lab brings together researchers from Columbia, U Maryland, USC, and NYU.
We created the Zebra‑CoT dataset to enable interleaved vision–language reasoning and have developed state-of-the-art visual reasoning models built on this foundation.
datasets 19
multimodal-reasoning-lab/Zebra-CoT
Viewer • Updated • 160k • 10.6k • 78
multimodal-reasoning-lab/Chess
Viewer • Updated • 20.5k • 528 • 6
multimodal-reasoning-lab/Connect-Four
Viewer • Updated • 2.03k • 117
multimodal-reasoning-lab/Ciphers
Viewer • Updated • 6.59k • 247
multimodal-reasoning-lab/Checkers
Viewer • Updated • 2.75k • 105 • 1
multimodal-reasoning-lab/Chemistry
Viewer • Updated • 4.67k • 462 • 11
multimodal-reasoning-lab/Visual-Jigsaw
Viewer • Updated • 21.9k • 572 • 4
multimodal-reasoning-lab/Physics
Viewer • Updated • 7.09k • 441 • 10
multimodal-reasoning-lab/RPM
Viewer • Updated • 3k • 103 • 1
multimodal-reasoning-lab/Maze
Viewer • Updated • 20k • 1.9k