anthracite-org/c2_logs_32k_llama3_qwen2_v1.3

这是一个包含角色扮演日志的数据集，采用ShareGPT格式。数据集已经过清洗，包括删除空白系统提示、长度过短的条目以及代理错误。数据集被缩短到32k的上下文长度，并修复了助手的回答中缺少空格的问题。This dataset consists of roleplay logs in ShareGPT format. The dataset has been cleaned, including the removal of blank system prompts, entries with insufficient length, and proxy errors. The dataset has been shortened to a 32k context length and fixes issues with missing spaces in the assistants responses.