brl-xfact/hum_eye4b

这是一个包含图像和对应标注信息的测试数据集，具体包括图像的标注ID、标注一致性、人工标注、简洁性、深度准确性、方向准确性、虚假信息标注、唯一标识符、图像路径、响应时间、总体评价、请求内容、安全性、充分性以及模型生成的标注。测试集共有2534个示例，数据集大小为3399419字节，下载大小为655958字节。

This is a test dataset containing images and corresponding annotation information, including annotation ID, annotation agreement, human captions, conciseness, depth accuracy, direction accuracy, hallucination annotations, unique identifiers, image paths, response times, overall ratings, request content, safety, sufficiency, and model-generated captions. The test set consists of 2534 examples, with a dataset size of 3,399,419 bytes and a download size of 655,958 bytes.