RAGEVALUATION-HJKMY/Qwen7b_ragbench_delucionqa_400row_mistake_added

该数据集包含了问答对及其相关上下文信息，以及针对回答的多种评价和标注信息。具体字段包括：问答对的标识、问题、文档、回答、支持回答的句子、不支持回答的句子、整体支持的说明、相关性说明、相关句子键、利用的句子键、相关性评分、利用度评分、完整性评分、错误数量、错误分布、句子级别的关键点、正确性评分、答案相似度、连贯性评分、上下文相关性评分、上下文利用度评分、参与度评分、事实准确性评分、事实正确性F1评分、关键点完整性评分、关键点虚构评分、关键点不相关评分、学习促进评分、拒绝准确性评分等。数据集分为训练集，大小为6254629字节，包含400个样本。

The dataset includes question-answer pairs with related context information and various evaluation and annotation information for the answers. Specific fields include: identification of question-answer pairs, question, document, answer, sentences supporting the answer, sentences not supporting the answer, overall support explanation, relevance explanation, relevant sentence keys, utilized sentence keys, relevance score, utilization score, completeness score, number of mistakes, mistake distribution, key points at the sentence level, correctness score, answer similarity, coherence score, context relevance score, context utilization score, engagement score, factual accuracy score, factual correctness F1 score, key point completeness score, key point hallucination score, key point irrelevant score, learning facilitation score, refusal accuracy score, etc. The dataset is split into a training set, which is 6254629 bytes in size and contains 400 samples.