CHATS-Lab/Verbalized-Sampling-Joke-Generation

Verbalized Sampling (VS)方法在创造性生成任务（如笑话生成）中提高了多样性，同时保持了幽默质量。这个数据集由CHATS-Lab创建，包含来自多个最先进的LLM对特定主题笑话请求的响应。数据集评估了创造性多样性、质量保持、模式崩溃缓解和分布覆盖等方面。数据集包含多种配置和分割，每种配置代表一种不同的提示方法，每种分割代表一个不同的语言模型。

This dataset demonstrates how Verbalized Sampling (VS) increases diversity in creative generation tasks, specifically joke generation, while maintaining humor quality. It includes diverse jokes from state-of-the-art LLMs in response to prompts requesting jokes about specific topics, evaluating creative diversity, quality maintenance, mode collapse mitigation, and distribution coverage. The dataset contains various configurations and splits, each representing a different prompting method and language model.