Interpretation and Assumption Control v01
What this dataset is
This dataset evaluates whether a system interprets medical images without inventing certainty when information is missing or ambiguous.
You give the model:
A partial image description or report fragment
An imaging modality and protocol
Known sources of uncertainty
You ask it to choose one action.
PROCEED
CLARIFY
REFUSE
Guessing is a failure.
Why this matters
Many medical imaging errors arise after acquisition.
Common failure… See the full description on the dataset page:
https://huggingface.co/datasets/ClarusC64/interpretation-assumption-control-medimg-v01.
