| --- |
| dataset_info: |
| features: |
| - name: prompt |
| dtype: string |
| - name: llama_response |
| dtype: string |
| - name: refused |
| dtype: int64 |
| - name: prompt_label |
| dtype: string |
| - name: source |
| dtype: string |
| - name: refusal_logprob |
| dtype: float64 |
| - name: feat_10000_direct_inquiries_and_tasks_harmless |
| dtype: float64 |
| - name: feat_10737_explaining_concepts_or_skills_harmless |
| dtype: float64 |
| - name: feat_1083_direct_problem_solving_queries_harmless |
| dtype: float64 |
| - name: feat_10878_directed_fictional_content_generation_harmful |
| dtype: float64 |
| - name: feat_10940_request_for_negative_group_stereotypes_harmful |
| dtype: float64 |
| - name: feat_11007_requests_for_sensitive_details_harmful |
| dtype: float64 |
| - name: feat_11223_metaphorical_aggression_for_positive_goals_harmless |
| dtype: float64 |
| - name: feat_11236_prompting_group_denigration_harmful |
| dtype: float64 |
| - name: feat_11382_factual_and_explanatory_content_requests_harmless |
| dtype: float64 |
| - name: feat_11430_harmful_prescriptive_medical_advice_harmful |
| dtype: float64 |
| - name: feat_11634_specific_information_content_requests_harmless |
| dtype: float64 |
| - name: feat_11761_specific_instructional_prompts_harmful |
| dtype: float64 |
| - name: feat_11861_direct_task_information_request_harmless |
| dtype: float64 |
| - name: feat_12102_structured_information_request_harmful |
| dtype: float64 |
| - name: feat_12232_specific_information_procedure_requests_harmless |
| dtype: float64 |
| - name: feat_1227_harmful_framing_or_instruction_harmful |
| dtype: float64 |
| - name: feat_1581_potentially_problematic_information_requests_harmful |
| dtype: float64 |
| - name: feat_1636_requesting_sensitive_private_data_harmful |
| dtype: float64 |
| - name: feat_1728_task_oriented_communication_processing_harmless |
| dtype: float64 |
| - name: feat_1742_general_assistant_tasks_harmless |
| dtype: float64 |
| - name: feat_1836_benign_informational_analytical_requests_harmless |
| dtype: float64 |
| - name: feat_2197_requests_for_exploitation_tactics_harmful |
| dtype: float64 |
| - name: feat_2827_language_requiring_scrutiny_harmless |
| dtype: float64 |
| - name: feat_286_corporate_contractual_financial_specifics_harmless |
| dtype: float64 |
| - name: feat_2945_harmless_procedural_technical_tasks_harmless |
| dtype: float64 |
| - name: feat_3183_instructional_text_operations_harmless |
| dtype: float64 |
| - name: feat_3248_crafting_deceptive_exploitative_content_harmful |
| dtype: float64 |
| - name: feat_3277_knowledge_and_explanation_queries_harmless |
| dtype: float64 |
| - name: feat_3368_guidance_for_self_improvement_understanding_harmless |
| dtype: float64 |
| - name: feat_3536_concise_information_extraction_harmless |
| dtype: float64 |
| - name: feat_3767_explicit_content_boundary_requests_harmless |
| dtype: float64 |
| - name: feat_3772_facilitating_misguidance_or_illicit_acts_harmful |
| dtype: float64 |
| - name: feat_3868_request_for_public_address_harmless |
| dtype: float64 |
| - name: feat_3895_direct_factual_retrieval_task_harmless |
| dtype: float64 |
| - name: feat_3915_harmful_instructional_content_harmful |
| dtype: float64 |
| - name: feat_4027_enumeration_and_list_prompts_harmless |
| dtype: float64 |
| - name: feat_4112_serious_factual_inquiry_harmless |
| dtype: float64 |
| - name: feat_4202_communication_&_information_security_harmless |
| dtype: float64 |
| - name: feat_4205_personal_difficulties_and_coping_harmless |
| dtype: float64 |
| - name: feat_4211_generating_group_based_discrimination_harmful |
| dtype: float64 |
| - name: feat_4319_specific,_direct_inquiry_instruction_harmless |
| dtype: float64 |
| - name: feat_4396_publicly_disclosed_personal_health_harmful |
| dtype: float64 |
| - name: feat_4528_organizational_risk_communications_harmful |
| dtype: float64 |
| - name: feat_4590_soliciting_discriminatory_narratives_harmless |
| dtype: float64 |
| - name: feat_4644_specific_information_retrieval_harmless |
| dtype: float64 |
| - name: feat_4991_conspiracy_theory_generation_promotion_harmful |
| dtype: float64 |
| - name: feat_5449_request_for_email_subject_lines_harmful |
| dtype: float64 |
| - name: feat_5505_email_subject_line_task_harmless |
| dtype: float64 |
| - name: feat_5509_simple_linguistic_task_prompts_harmless |
| dtype: float64 |
| - name: feat_573_specific_information_or_action_request_harmful |
| dtype: float64 |
| - name: feat_6090_information_management_and_analysis_harmless |
| dtype: float64 |
| - name: feat_6238_factual_definition_queries_harmless |
| dtype: float64 |
| - name: feat_6469_academic_advocacy_content_task_harmless |
| dtype: float64 |
| - name: feat_6499_constructive_guidance_seeking_harmless |
| dtype: float64 |
| - name: feat_6562_direct_task_information_seeking_harmless |
| dtype: float64 |
| - name: feat_6733_content_&_information_integrity_harmful |
| dtype: float64 |
| - name: feat_6815_generating_influential_content_harmful |
| dtype: float64 |
| - name: feat_7063_directive_explanations_&_creations_harmless |
| dtype: float64 |
| - name: feat_7100_requests_for_specific_actionable_data_harmful |
| dtype: float64 |
| - name: feat_7261_seeking_safety_and_protection_guidance_harmless |
| dtype: float64 |
| - name: feat_7294_propagating_harmful_or_fraudulent_actions_harmful |
| dtype: float64 |
| - name: feat_7480_email_subject_line_query_harmless |
| dtype: float64 |
| - name: feat_7565_instructional_content_generation_harmless |
| dtype: float64 |
| - name: feat_7575_explicit_information_requests_harmless |
| dtype: float64 |
| - name: feat_7677_constrained_text_generation_instruction_harmless |
| dtype: float64 |
| - name: feat_7897_action_oriented_practical_queries_harmless |
| dtype: float64 |
| - name: feat_7940_strategic_deception_or_corporate_manipulation_harmful |
| dtype: float64 |
| - name: feat_7971_factual_entity_information_queries_harmless |
| dtype: float64 |
| - name: feat_8063_noteworthy_entities'_positive_ethical_aspects_harmless |
| dtype: float64 |
| - name: feat_808_promoting_group_discrimination_by_traits_harmful |
| dtype: float64 |
| - name: feat_8193_detailed_content_generation_requests_harmful |
| dtype: float64 |
| - name: feat_823_instructions_for_mathematical_tasks_harmless |
| dtype: float64 |
| - name: feat_8240_explicit_structured_information_tasks_harmless |
| dtype: float64 |
| - name: feat_8643_neutral_framed_content_tasks_harmless |
| dtype: float64 |
| - name: feat_9081_managing_sensitive_critical_information_harmful |
| dtype: float64 |
| - name: feat_910_email_subject_line_task_harmless |
| dtype: float64 |
| - name: feat_9267_contextually_harmless_fictional_scenarios_harmless |
| dtype: float64 |
| - name: feat_9623_disinformation_for_manipulation_harmful |
| dtype: float64 |
| - name: feat_9649_**soliciting_harmful_content_creation**_harmful |
| dtype: float64 |
| - name: feat_9994_requests_for_n_item_lists_harmless |
| dtype: float64 |
| splits: |
| - name: train |
| num_bytes: 7048606 |
| num_examples: 5500 |
| download_size: 4181665 |
| dataset_size: 7048606 |
| configs: |
| - config_name: default |
| data_files: |
| - split: train |
| path: data/train-* |
| --- |
| |
