Dataset Viewer
Auto-converted to Parquet Duplicate
image
imagewidth (px)
32
32
concepts
list
label
class label
2 classes
label_name
stringclasses
2 values
[ 1, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0 ]
1glorp
glorp
[ 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0 ]
0drent
drent
[ 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 1, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1 ]
1glorp
glorp
[ 1, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0 ]
0drent
drent
[ 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1 ]
1glorp
glorp
[ 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 1, 0, 1, 1, 0, 1, 0, 0, 0, 0 ]
0drent
drent
[ 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0 ]
1glorp
glorp
[ 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0 ]
1glorp
glorp
[ 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1 ]
1glorp
glorp
[ 0, 1, 1, 1, 0, 1, 0, 0, 0, 1, 0, 0 ]
1glorp
glorp
[ 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0 ]
0drent
drent
[ 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0 ]
0drent
drent
[ 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0 ]
1glorp
glorp
[ 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0 ]
0drent
drent
[ 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1 ]
1glorp
glorp
[ 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0 ]
1glorp
glorp
[ 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 1, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0 ]
1glorp
glorp
[ 0, 1, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0 ]
0drent
drent
[ 1, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0 ]
0drent
drent
[ 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
0drent
drent
[ 1, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0 ]
0drent
drent
[ 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
1glorp
glorp
[ 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 1, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 1, 1, 0, 0, 1, 0, 1, 0, 0, 0, 0 ]
0drent
drent
[ 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0 ]
1glorp
glorp
[ 1, 1, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0 ]
1glorp
glorp
[ 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 1, 0 ]
1glorp
glorp
[ 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1 ]
1glorp
glorp
[ 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
1glorp
glorp
[ 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1 ]
1glorp
glorp
[ 1, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0 ]
1glorp
glorp
[ 0, 0, 1, 0, 1, 1, 0, 0, 1, 0, 0, 0 ]
0drent
drent
[ 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1 ]
1glorp
glorp
[ 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0 ]
1glorp
glorp
[ 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0 ]
1glorp
glorp
[ 0, 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0 ]
0drent
drent
[ 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0 ]
1glorp
glorp
[ 1, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0 ]
1glorp
glorp
[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0 ]
1glorp
glorp
[ 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0 ]
1glorp
glorp
[ 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0 ]
1glorp
glorp
[ 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1 ]
1glorp
glorp
[ 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0 ]
1glorp
glorp
End of preview. Expand in Data Studio

Robots — Human Concepts

Synthetic benchmark for evaluating Concept Bottleneck Models (CBMs) under finer-grained, human-annotated concepts. Same underlying robot images and labels as juliannski/robots-true-concepts, but the foot_shape ground-truth concept is replaced by 6 one-hot subtypes that a human annotator would actually see, modelling concept specification mismatch between annotators and the latent labeling rule.

Generated from

This dataset is the exact output of the concept-benchmark Python package — seed and config are pinned for bit-identical reproduction.

# pip install concept-benchmark==0.3.1
from concept_benchmark.robots import DatasetGenerator
dataset = DatasetGenerator(seed=1014, concept_preset="foot_subtypes").generate()

📓 Quickstart notebook — load from the Hub → train a CBM → run oracle interventions, end-to-end in one notebook.

🔗 PyPI · GitHub

Quick start (datasets library)

from datasets import load_dataset
ds = load_dataset("juliannski/robots-human-concepts")
row = ds["train"][0]
row["image"]       # PIL.Image (32x32 RGBA)
row["concepts"]    # 12-dim binary list, ordered by concept_names below
row["label"]       # 0 or 1
row["label_name"]  # "drent" or "glorp"

Schema

Field Type Notes
image datasets.Image() 32x32 RGBA PNG, embedded
concepts Sequence(int8, length=12) binary; index order = concept_names below
label ClassLabel("drent", "glorp")
label_name string display alias for label

concept_names (column order in concepts):

["head_shape", "body_shape", "has_knees", "has_antennae",
 "ears_shape", "mouth_type",
 "foot_shape_flat_trapezoid", "foot_shape_flat_square", "foot_shape_flat_5sided",
 "foot_shape_pointy_rounded", "foot_shape_pointy_square", "foot_shape_pointy_4sided"]

Splits

Split n
train 16,576
validation 4,144
test 10,000

Stratified on label; seed=1014. Class distribution intentionally imbalanced (~87% glorp).

Companion datasets

  • juliannski/robots-true-concepts — same robots, 7 ground-truth concepts (foot_shape as a single binary pointy/flat).
  • juliannski/sudoku — sudoku validation benchmark.

License

MIT.

Downloads last month
65