InternScience/SGI-DryExperiment
收藏Hugging Face2025-12-30 更新2026-02-07 收录
下载链接：
https://hf-mirror.com/datasets/InternScience/SGI-DryExperiment
下载链接
链接失效反馈官方服务：
资源简介：
SGI-Bench是一个用于评估科学通用智能（SGI）的数据集，涵盖了科学探究的完整循环：思考、构思、行动和感知。该数据集跨越10个学科，包含超过1,000个专家策划的样本，灵感来源于《科学》杂志的125个重大问题。数据集包括科学深度研究、想法生成、干/湿实验和多模态实验推理等任务。数据集通过专家评审和数据清理确保高质量和挑战性。
SGI-Bench is a dataset designed to evaluate Scientific General Intelligence (SGI) across the full inquiry cycle: Deliberation, Conception, Action, and Perception. The benchmark spans 10 disciplines and includes more than 1,000 expert-curated samples inspired by Science’s 125 Big Questions. It features tasks such as scientific deep research, idea generation, dry/wet experiments, and multimodal experimental reasoning. The dataset ensures high fidelity and challenge through expert review and data cleaning.
提供机构：
InternScience
