Safety
AEGIS 2.0
Assesses the safety of model responses using human and model annotations of harmful content.
11,500items
1subjects
CC-BY-4.0license
safetydomain
textmodality
item-level responses released
Saturation status: No
About this data
Dataset name: AEGIS 2.0 response safety
Released Mistral-7B-v0.1 completions and their native response safety annotations. Prompt-only classification examples and synthetic refusal augmentation remain in the captured source.
11,500 items · 11,608 recorded attempts. Browse item pages to see all observations. Filters and means apply to the current page.
Loading observation records…