s the collection of real-world network IDS alerts that contain 25 types of attacks and payload information. The dataset has 153,829 IDS alerts which are labeled as malicious(label:1) and benign(label:2) respectively. Also, all of the words in payload are embedded as a word vector to 100-dimension for each word. The purpose of the dataset is to encourage IDS research in terms of real-world environments considering payload contents. It is divided into training (80%) and test set (20%) with 4-columned TAB separated structure (idx, category, wordvector(payload), label).