dataset_info:
features:
- name: prompt
dtype: string - name: response
dtype: string - name: category
dtype: string - name: source
dtype: string - name: judge_score
struct:- name: prompt_is_harmful
dtype: float64 - name: response_is_refusal_or_dodge
dtype: float64 - name: response_contains_substantive_harm
dtype: float64 - name: jailbreak_label
dtype: float64 - name: rationale
dtype: string - name: source
dtype: string
- name: prompt_is_harmful
- name: label
dtype: int64 - name: activation_layer_4
list: float64 - name: activation_layer_10
list: float64 - name: activation_layer_15
list: float64 - name: activation_layer_20
list: float64 - name: activation_layer_25
list: float64 - name: activation_layer_30
list: float64
splits: - name: train
num_bytes: 851310228
num_examples: 4315 - name: validation
num_bytes: 121533058
num_examples: 616 - name: test
num_bytes: 243456010
num_examples: 1234
download_size: 1214667407
dataset_size: 1216299296
configs: - config_name: default
data_files:- split: train
path: data/train-* - split: validation
path: data/validation-* - split: test
path: data/test-*
- split: train