Zebra-CoT: A Dataset for Interleaved Vision Language Reasoning

TL;DR

Zebra-CoT introduces a large-scale dataset with interleaved text-image reasoning traces to improve multimodal reasoning in AI models, addressing current performance and data scarcity challenges.

Contribution

The paper presents Zebra-CoT, a diverse dataset of over 180,000 samples for training visual chain of thought models, enabling significant performance improvements.

Findings

+12% accuracy on test set after fine-tuning

Up to +13% gain on VLM benchmarks

High-quality visual reasoning chains generated

Abstract

Humans often use visual aids, for example diagrams or sketches, when solving complex problems. Training multimodal models to do the same, known as Visual Chain of Thought (Visual CoT), is challenging due to: (1) poor off-the-shelf visual CoT performance, which hinders reinforcement learning, and (2) the lack of high-quality visual CoT training data. We introduce $\textbf{Zebra-CoT}$, a diverse large-scale dataset with 182,384 samples, containing logically coherent interleaved text-image reasoning traces. We focus on four categories of tasks where sketching or visual reasoning is especially natural, spanning scientific questions such as geometry, physics, and algorithms; 2D visual reasoning tasks like visual search and jigsaw puzzles; 3D reasoning tasks including 3D multi-hop inference, embodied and robot planning; visual logic problems and strategic games like chess. Fine-tuning the Anole-7B model on the Zebra-CoT training corpus results in an improvement of +12% in our test-set accuracy and yields up to +13% performance gain on standard VLM benchmark evaluations. Fine-tuning Bagel-7B yields a model that generates high-quality interleaved visual reasoning chains, underscoring Zebra-CoT's effectiveness for developing multimodal reasoning abilities. We open-source our dataset and models to support development and evaluation of visual CoT.