Salesforce/shared-imagination

Shared Imagination数据集包含了用于TMLR论文《Shared Imagination: LLMs Hallucinate Alike》中的问题。该数据集有两个配置：默认配置（default）和无选项打乱配置（no_choice_shuffle）。每个配置下有四个数据split：直接问题（direct_questions）、上下文问题（context_questions）、创造性直接问题（direct_questions_creative）和创造性上下文问题（context_questions_creative）。数据集包含了由不同模型生成的问题，涵盖数学、计算机科学、物理、化学、生物、地理、社会学、心理学、经济学、会计、市场营销、法律、政治、历史、文学、哲学和宗教等领域。此外，还有关于创造性写作故事的问题。每个问题实例包括模型、类别、索引、问题文本、选项列表、正确选项索引、知识段落或故事（上下文）以及概念等信息。

The Shared Imagination dataset contains the questions used in the TMLR paper Shared Imagination: LLMs Hallucinate Alike. There are two configurations: the default config and the no_choice_shuffle config. Under each config, there are four splits: direct_questions, context_questions, direct_questions_creative, and context_questions_creative. The dataset includes questions generated by different models, covering subjects such as mathematics, computer science, physics, chemistry, biology, geography, sociology, psychology, economics, accounting, marketing, law, politics, history, literature, philosophy, and religion, as well as questions about creative writing stories. Each question instance includes the model, category, index, question text, list of choices, index of the correct choice, knowledge paragraph or story (context), and concept information.