rlhn/remove-400K
收藏Hugging Face2025-05-27 更新2025-11-01 收录
下载链接：
https://hf-mirror.com/datasets/rlhn/remove-400K
下载链接
链接失效反馈官方服务：
资源简介：
RLHN是一个级联的大型语言模型框架，旨在准确重标记现有IR/RAG训练数据集中的难负样本，例如MS MARCO和HotpotQA。这个Tevatron数据集（400K训练对）包含原始查询、正例和删除了每个训练对中的单个错误负例后的难负例。该数据集可用于精炼嵌入、ColBERT或多元向量以及重排模型。
RLHN is a cascading LLM framework designed to accurately relabel hard negatives in existing IR/RAG training datasets, such as MS MARCO and HotpotQA. This Tevatron dataset (400K training pairs) contains the original queries, positives, and hard negatives after dropping each training pair with a single false negative. The dataset is used for fine-tuning embedding, ColBERT, or multi-vector and reranker models.
提供机构：
rlhn
