BatsResearch/sycl

SyCL数据集是用于研究实验的合成数据集，基于MS MARCO查询数据生成，包含了由Llama 3.3 70B、Qwen2.5 72B和Qwen2.5 32B模型生成的数据。数据集具有多个相关性标注级别，并提供了真实的段落数据，包括BM25检索的段落、TREC DL 2020查询和对应的注释。The SyCL dataset is a synthetic dataset for research experiments, based on MS MARCO query data, containing data generated by Llama 3.3 70B, Qwen2.5 72B, and Qwen2.5 32B models. The dataset includes multiple levels of relevance annotations and provides real passage data, including BM25 retrieved passages, TREC DL 2020 queries, and their annotations.