license: mit
task_categories:
- image-to-text
- text-to-image
language:
- en
pretty_name: simons ARC (abstraction & reasoning corpus) solve gravity version 14
size_categories:
- 1K<n<10K
configs:
- config_name: default
data_files:
- split: train
path: data.jsonl
Version 1
ARC-AGI Tasks where the job is to apply gravity in the directions up/down/left/right.
example count: 2-4.
test count: 1-2.
image size: 3-10.
number of pixels to apply gravity to: 2-5.
Exercises image_gravity_move().
Version 2
Exercises image_gravity_draw().
Version 3
Exercises image_gravity_move() and image_gravity_draw().
Increased max_number_of_positions from 5 to 8.
Version 4
image size: 3-30.
max number of positions: 5.
Version 5
image size: 3-30.
max number of positions: 7.
Version 6
Added diagonal movement with image_gravity_move().
image size: 3-10.
max number of positions: 7.
Version 7
Focus on diagonal movement with image_gravity_move().
image size: 3-20.
max number of positions: 5.
Version 8
Focus on diagonal movement with image_gravity_move().
image size: 3-30.
max number of positions: 5.
Version 9
Exercises image_gravity_move() and image_gravity_draw().
image size: 3-30.
max number of positions: 7.
Version 10
Focus on diagonal movement with image_gravity_draw().
image size: 3-10.
max number of positions: 5.
Version 11
Focus on diagonal movement with image_gravity_draw().
image size: 3-20.
max number of positions: 5.
Version 12
Focus on diagonal movement with image_gravity_draw().
image size: 3-30.
max number of positions: 5.
Version 13
Earlier predictions added to some of the rows.
Version 14
Added fields: arc_task, test_index, earlier_output.