metadata
dataset_info:
features:
- name: url
dtype: string
- name: task_id
dtype: int64
- name: run_id
dtype: int64
- name: step_idx
dtype: int64
- name: objective
dtype: string
- name: previous_actions
sequence: string
- name: observation
dtype: string
- name: current_action
dtype: string
- name: next_state_acctree
dtype: string
- name: raw_prediction
dtype: string
- name: parsed_rationale
dtype: string
- name: parsed_next_state
dtype: string
splits:
- name: train
num_bytes: 181993520
num_examples: 13494
download_size: 37533102
dataset_size: 181993520
configs:
- config_name: default
data_files:
- split: train
path: data/train-*