官方服务:
资源简介:
HalluQA包含450个手工构造的问答题目,分为误导类问题和知识类问题两大类,可以用来评估中文大型语言模型及对话系统中的幻觉现象。
应用场景:
提供机构:
chengqinyuan创建时间:
2023-11-09
相关数据集
CohenQu/POPE-93-medium-no_guide-easy
该数据集包含训练集和测试集,训练集有512个样本,测试集有180个样本。数据集特征包括数据来源(data_source)、提示(prompt,包含内容和角色)、能力(ability)、奖励模型(reward_model,包含真实结果和风格)、额外信息(extra_info,包含索引和分割)、等级(level)和问题ID(problem_id)。数据集的总下载大小为266610字节,总大小为2791
Hugging Face2025-12-13 更新150
Feudor2/HE_hallucination_detection
--- dataset_info: features: - name: messages list: - name: content dtype: string - name: role dtype: string splits: - name: train num_bytes: 4771061 num_example
Hugging Face2026-03-26 更新80
dmaniloff/glassbox_ragtruth_hallucination_test
--- dataset_info: features: - name: prompt dtype: string - name: response dtype: string - name: label dtype: int64 - name: unique_id dtype: string - name: failure_mode
Hugging Face2026-03-24 更新50
ai-safety-institute/qwen3_5_27b_ab_hallucinates_citations_rollouts
--- dataset_info: features: - name: instruction dtype: string - name: reasoning dtype: string - name: output dtype: string - name: reasoning_pre_rewrite dtype: string - nam
Hugging Face2026-04-30 更新90
dmaniloff/glassbox_halubench_hallucination_test
--- dataset_info: features: - name: prompt dtype: string - name: response dtype: string - name: label dtype: int64 - name: unique_id dtype: string - name: failure_mode
Hugging Face2026-03-24 更新60



