H-ARC (human-ARC)
收藏资源简介:
H-ARC数据集由纽约大学心理学系和数据科学中心创建,旨在评估人类在抽象推理和推理任务中的表现。该数据集包含15744次人类尝试解决ARC基准测试中的400个训练任务和400个评估任务的记录,包括详细的步骤和动作追踪。数据集的创建过程涉及招募1729名参与者,通过Amazon Mechanical Turk平台进行实验。H-ARC数据集的应用领域主要是人工智能和认知科学,旨在通过对比人类和机器的表现,推动更智能和类人系统的开发。
The H-ARC dataset was developed by the Department of Psychology and the Center for Data Science at New York University, with the objective of evaluating human performance on abstract reasoning and reasoning tasks. This dataset comprises 15,744 records of human attempts to solve 400 training tasks and 400 evaluation tasks from the ARC benchmark, including detailed step-by-step processes and action tracking data. The creation of this dataset involved recruiting 1,729 participants to carry out experiments through the Amazon Mechanical Turk platform. The H-ARC dataset is primarily applied in the fields of artificial intelligence and cognitive science, aiming to advance the development of more intelligent and human-like systems by comparing the performance between humans and machines.




