metadata
dataset_info:
features:
- name: experiment_id
dtype: int64
- name: step
dtype: int64
- name: sequence
dtype: string
- name: activation
dtype: float64
- name: metric_value
dtype: float64
- name: label
dtype: string
- name: pos_match_string
dtype: string
- name: neg_match_string
dtype: string
splits:
- name: train
num_bytes: 206536
num_examples: 1208
download_size: 35834
dataset_size: 206536
configs:
- config_name: default
data_files:
- split: train
path: data/train-*