ai2-adapt-dev/tulu_v3.9_wildchat_unused

该数据集包含了详细的对话信息，包括对话的哈希值、使用的模型、时间戳、对话内容、语言、国家、IP地址哈希、用户代理信息、消息内容、角色、状态、是否经过编辑、是否含有毒性内容等。此外，数据集还包括了OpenAI和Detoxify的审核结果，如骚扰、仇恨、自残、性内容、暴力等类别的分类和评分。数据集分为训练集，包含140,860个样本，总大小为3,393,823,428.1167483字节。

The dataset contains detailed conversation information, including conversation hash, model used, timestamp, conversation content, language, country, hashed IP address, user agent information, message content, role, state, whether redacted, and whether it contains toxic content. Additionally, the dataset includes moderation results from OpenAI and Detoxify, such as classifications and scores for categories like harassment, hate, self-harm, sexual content, and violence. The dataset is divided into a training set, containing 140,860 samples, with a total size of 3,393,823,428.1167483 bytes.