vanta-research/reasoned-refusal

Reasoned Refusal数据集是一个精心策划的数据集，包含1,400个对话示例，展示了如何拒绝无益、误导或适得其反的请求，同时解释原因并提供建设性的替代方案。该数据集旨在通过教导模型何时以及如何拒绝请求来微调语言模型，使其真正变得有帮助。数据集的关键特点包括解释拒绝的原因、识别实际需求、提供替代方案、保持尊重但坚定的态度以及详细的推理。数据集的结构为JSONL格式，包含用户消息和助理响应的对话对。

The Reasoned Refusal dataset is a curated dataset of 1,400 conversational examples demonstrating how to decline unhelpful, misguided, or counterproductive requests while explaining the reasoning and offering constructive alternatives. Designed for fine-tuning language models to be genuinely helpful by knowing when and how to say no. Key characteristics include explaining the why, identifying actual needs, offering alternatives, maintaining respectful but firm boundaries, and providing detailed reasoning. The dataset is structured in JSONL format with user messages and assistant responses.