-
AI Safety Dataset
The dataset used for bibliometric analysis of the literature on AI safety. -
SciMT-Safety
The SciMT-Safety dataset is a benchmark for evaluating the safety of AI systems in science. It consists of hundreds of refined red-teaming queries that span the fields of... -
Anthropic red-team dataset
The Anthropic red-team dataset is a significant open-access dataset aimed at improving AI safety through training preference models and assessing their safety.