dataset_info:
features:
- name: Index
dtype: int64 - name: goal
dtype: string - name: prompt
dtype: string - name: target_response
dtype: string - name: human1
dtype: int64 - name: human2
dtype: int64 - name: human3
dtype: int64 - name: human_majority
dtype: int64 - name: harmbench_cf
dtype: int64 - name: gpt4_cf
dtype: int64 - name: llamaguard2_cf
dtype: int64 - name: llama3_cf
dtype: int64 - name: embedding
sequence: float32
splits: - name: test
num_bytes: 1301579
num_examples: 300
download_size: 1284511
dataset_size: 1301579
configs: - config_name: default
data_files:- split: test
path: data/test-*
- split: test