metadata
dataset_info:
features:
- name: statement
dtype: string
- name: label
dtype: bool
splits:
- name: train
num_bytes: 91207
num_examples: 1200
download_size: 33835
dataset_size: 91207
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: apache-2.0
pretty_name: True-False Questions dataset - Facts about companies
The dataset's authors are Amos Azaria and Tom Mitchell, which they first used in their The Internal State of an LLM Knows When It's Lying paper