llm-semantic-router/fact-check-classification-dataset

该数据集是一个用于二分类的数据集，旨在训练分类器以判断一个提示是否需要外部事实核查。数据集包含两个主要标签：FACT_CHECK_NEEDED（需要外部事实核查）和NO_FACT_CHECK_NEEDED（不需要外部事实核查）。数据集来源于多个公开数据集，包括NISQ、FactCHD、FaithDial、HaluEval、SQuAD、TruthfulQA、TriviaQA、HotpotQA、RAG Dataset等。数据集结构包括文本、标签、标签ID、来源和原始标签等字段。

This dataset is a binary classification dataset designed to train classifiers to determine whether a prompt requires external fact-checking. It includes two main labels: FACT_CHECK_NEEDED (requires external fact-checking) and NO_FACT_CHECK_NEEDED (does not require external fact-checking). The dataset is sourced from multiple public datasets, including NISQ, FactCHD, FaithDial, HaluEval, SQuAD, TruthfulQA, TriviaQA, HotpotQA, RAG Dataset, etc. The dataset structure includes fields such as text, label, label_id, source, and original_label.