metadata
dataset_info:
features:
- name: UID
dtype: string
- name: Question
dtype: string
- name: Answer
dtype: string
- name: Type
dtype: string
- name: PatientID
dtype: string
- name: Age
dtype: int64
- name: HeartSize
dtype: int64
- name: PulmonaryCongestion
dtype: int64
- name: PleuralEffusion_Right
dtype: int64
- name: PleuralEffusion_Left
dtype: int64
- name: PulmonaryOpacities_Right
dtype: int64
- name: PulmonaryOpacities_Left
dtype: int64
- name: Atelectasis_Right
dtype: int64
- name: Atelectasis_Left
dtype: int64
- name: Split
dtype: string
- name: PhysicianID
dtype: string
- name: StudyDate
dtype: string
- name: Sex
dtype: string
- name: Image
dtype: image
splits:
- name: train
num_bytes: 5622656901
num_examples: 20288
- name: val
num_bytes: 1462315894
num_examples: 5120
- name: test
num_bytes: 1783934753
num_examples: 6592
download_size: 363809891
dataset_size: 8868907548
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: val
path: data/val-*
- split: test
path: data/test-*
TAIX-VQA Dataset
We share this dataset used for the evaluations in "Evaluating Reasoning Faithfulness in Medical Vision-Language Models using Multimodal Perturbations". The dataset contains 1,000 distinct chest X-rays from the TAIX-RAY dataset with structured annotations from human radiologists. For each image, we added 32 different, clinically realistic questions together with expert-annotated answers.
Dataset Details
For details, please check the paper, project page, and Github.
✏️ Citation
If you find this work useful, please cite:
@article{evaluating-2025,
title={Evaluating Reasoning Faithfulness in Medical Vision-Language Models using Multimodal Perturbations},
author={Moll, Johannes and Graf, Markus and Lemke, Tristan and Lenhart, Nicolas and Truhn, Daniel and Delbrouck, Jean-Benoit and Pan, Jiazhen and Rueckert, Daniel and Adams, Lisa C. and Bressem, Keno K.},
journal={arXiv preprint arXiv:2510.11196},
url={https://arxiv.org/abs/2510.11196},
year={2025}
}