Attention Judgment Task Data

该数据集包含了在视觉分类任务和注意力判断任务上训练人工代理的数据，旨在评估注意力架构对代理效果的影响。数据集包括在两种角色——发送者和接收者代理中使用的注意力激活值，并基于注意力架构的存在采用了2×2的设计进行训练。其中，注意力判断任务旨在评估一个代理对另一个代理注意力状态进行分类判断的能力。

This dataset contains training data for artificial agents on visual classification tasks and attention judgment tasks, with the objective of evaluating the impact of attention architectures on agent performance. The dataset includes attention activation values used by two agent roles: sender and receiver agents, and these agents were trained using a 2×2 experimental design based on the presence or absence of attention architectures. Specifically, the attention judgment task is designed to assess an agent's capacity to perform classification-based judgment on the attention states of another agent.