tasksource/synthetic-retrieval-NLI
收藏Hugging Face2024-12-23 更新2025-04-12 收录
下载链接:
https://hf-mirror.com/datasets/tasksource/synthetic-retrieval-NLI
下载链接
链接失效反馈官方服务:
资源简介:
这是一个设计用于探测文本编码器中长文本上下文检索能力的合成NLI任务数据集,包含binary、count和position三种配置。每种配置都包括前提(premise)、假设(hypothesis)、标签(label)、词汇表(vocab)和n字段。数据集分为训练集、验证集和测试集。
This is a synthetic NLI task dataset designed to probe long context retrieval in text encoders, including three configurations: binary, count, and position. Each configuration includes fields for premise, hypothesis, label, vocabulary, and n. The dataset is split into training, validation, and test sets.
提供机构:
tasksource



