README.md

---
dataset_info:
  features:
  - name: object
    dtype: string
  - name: singular_object
    dtype: string
  - name: plural_object
    dtype: string
  - name: correct_answer
    dtype: string
  - name: incorrect_answer
    dtype: string
  - name: original_correct_answers
    list: string
  - name: color_entropy
    dtype: float64
  - name: color_probs
    struct:
    - name: white
      dtype: float64
    - name: black
      dtype: float64
    - name: brown
      dtype: float64
    - name: green
      dtype: float64
    - name: grey
      dtype: float64
    - name: red
      dtype: float64
    - name: silver
      dtype: float64
    - name: yellow
      dtype: float64
    - name: orange
      dtype: float64
    - name: blue
      dtype: float64
    - name: pink
      dtype: float64
    - name: gold
      dtype: float64
    - name: brass
      dtype: float64
    - name: tan
      dtype: float64
    - name: burgundy
      dtype: float64
    - name: purple
      dtype: float64
  - name: top_colors
    list: string
  - name: argmax_color
    dtype: string
  - name: original_image
    dtype: image
  - name: counterfact_image
    dtype: image
  - name: entropy_category
    dtype: string
  - name: entropy_rank
    dtype: int64
  splits:
  - name: train
    num_bytes: 229565456
    num_examples: 493
  download_size: 226599863
  dataset_size: 229565456
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---