datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
measurement-db
The AI Measurement Data Bank
Measurement Data Bank is a curated collection of standardized, item-level AI
evaluation results for measurement-science analysis.
Schema and versioned downloads
These six benchmark datasets use schema version 3. The GitHub repository contains the corresponding builders, source manifests, tests, and curation records.
Grading criteria are stored in items.parquet; traces link to observations through response_id. The response table on… See the full description on the dataset page: https://huggingface.co/datasets/aims-foundations/measurement-db.moral_stories_foundations
Moral Stories, labelled by moral foundation
Why moral foundations? Moral Foundations Theory is one of the few maps of human values that is
calibrated against real people: it was drawn from cross-cultural survey studies and aims to hold
across societies, not just Western ones. That breadth is what matters here. If we want to study
how a constructed intelligence, a kind of moral alien, reasons about right and wrong, we need a
measure that generalises beyond any one culture. Moral… See the full description on the dataset page: https://huggingface.co/datasets/wassname/moral_stories_foundations.openart-painterly-foundations
OpenArt — Painterly Foundations
openart-painterly-foundations is the painterly foundations subject collection of the OpenArt family
of open, public-domain art datasets: 13,786 works (9,107 paintings/illustrations · 4,596
photographed objects · 83 unclassified), each paired with a structured VLM caption plus
medium, attribution and inscription metadata.
The painting-forward core of OpenArt — the 2-D fine-art techniques that anchor the openbrush brand: drawing, etching, oil… See the full description on the dataset page: https://huggingface.co/datasets/jaddai/openart-painterly-foundations.