metadata
dataset_info:
features:
- name: anchor_text
dtype: string
- name: source_dataset
dtype: string
- name: source_index
dtype: float64
- name: anchor_length
dtype: int64
- name: paraphrase_text
dtype: string
- name: paraphrase_length
dtype: int64
- name: generation_strategy
dtype: string
- name: generation_model
dtype: string
- name: llm_score_claude-sonnet-4-6
dtype: float64
- name: human_score_Lee
dtype: int64
splits:
- name: train
num_bytes: 17846
num_examples: 67
download_size: 12131
dataset_size: 17846
configs:
- config_name: default
data_files:
- split: train
path: data/train-*