UCSC-VLAA/STAR-benign-915
收藏官方服务：
资源简介：
STAR-1是一个为了提高大型推理模型（LRMs）安全性对齐而设计的高质量安全数据集。该数据集基于多样性、深思熟虑的推理和严格的过滤原则，整合并优化了多个来源的数据，提供了以政策为基础的推理样本。数据集中包含1000个经过精心挑选的示例，每个示例都通过GPT-4o-based评估与最佳安全实践保持一致。
STAR-1 is a high-quality safety dataset designed to enhance safety alignment in large reasoning models (LRMs). The dataset is built on the principles of diversity, deliberative reasoning, and rigorous filtering, integrating and refining data from multiple sources to provide policy-grounded reasoning samples. It contains 1,000 carefully selected examples, each aligned with best safety practices through GPT-4o-based evaluation.
提供机构：
UCSC-VLAA
