---
dataset_info:
  features:
  - name: prompt
    dtype: string
  - name: strat
    dtype: string
  - name: generated_text
    dtype: string
  - name: Bayesian_surprise
    dtype: float64
  - name: Good_idx
    dtype: float64
  - name: Howard_idx
    dtype: float64
  - name: M_ER_gap
    dtype: float32
  - name: C_ER_gap
    dtype: float32
  - name: diversity
    dtype: float64
  - name: quality_score
    dtype: float64
  - name: novelty
    dtype: float64
  splits:
  - name: train
    num_bytes: 18358
    num_examples: 16
  download_size: 19040
  dataset_size: 18358
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---