Beta
Explore
Marketplace
Neural Labs
Chat
Wallet
Docs
RAH-Bench – Dataset by MM-Hallu | AlphaNeural AI
You can deploy this model and start earning money today!
MM-Hallu
/
RAH-Bench
like
0
visual-question-answering
en
apache-2.0
1K<n<10K
parquet
image
text
datasets
dask
polars
mlcroissant
2311.16479
us
hallucination-evaluation
object-hallucination
multimodal
vision-language-model
COCO
Views
No views yet
Model card
Files and Versions
Community
API
RAH-Bench
Benchmark for evaluating object hallucination in VLMs. 3,000 binary yes/no questions about COCO val2017 images, categorized by hallucination type.
Fields
Field Description
image COCO val2017 image
question_id Unique question ID (1-3000)
coco_image_id COCO image ID
question Yes/no question about the image
label Ground truth: "yes" or "no"
type Hallucination category
Question Categories
type label count
attribute… See the full description on the dataset page:
https://huggingface.co/datasets/MM-Hallu/RAH-Bench
.
