dworsleytonks/project14-llm-processed-dataset

Project 14 – Medical LLM Processed Dataset是一个多语言（英语和法语）数据集，包含两种配置：dpo和sft。dpo配置包含提示、选择的和拒绝的响应、语言、来源和任务类型，分为训练和验证集。sft配置包含指令、响应、语言、来源、症状、前因、发现、置信度、任务类型和问题类型，分为训练、验证和测试集。数据集适用于文本生成和强化学习任务，采用cc-by-nc-4.0许可证。

Project 14 – Medical LLM Processed Dataset is a multilingual (English and French) dataset with two configurations: dpo and sft. The dpo configuration includes prompts, chosen and rejected responses, language, source, and task type, with train and validation splits. The sft configuration includes instructions, responses, language, source, symptoms, antecedents, findings, confidence, task type, and qtype, with train, validation, and test splits. The dataset is suitable for text-generation and reinforcement-learning tasks and is licensed under cc-by-nc-4.0.