osieosie/tulu-wildchat-filtered-mia-v4

该数据集包含一系列对话记录，每条记录包含对话的哈希值、使用的模型、时间戳、对话内容、用户国家、哈希IP、语言信息、是否经过编辑、对话角色、状态标记、是否含有毒性内容等。此外，还包含OpenAI和Detoxify对对话内容的审核结果，如各种毒性类别的标记和分数。数据集主要用于分析和研究对话中的毒性内容及其审核结果。

This dataset contains a series of conversation records, each including the conversation hash, model used, timestamp, conversation content, user country, hashed IP, language information, whether it has been redacted, conversation role, state markers, whether it contains toxic content, etc. Additionally, it includes moderation results from OpenAI and Detoxify, such as flags and scores for various toxicity categories. The dataset is primarily used for analyzing and researching toxic content in conversations and its moderation outcomes.