PawelPikulik/psnprofiles-trophy-guides

该数据集包含从PSNProfiles.com网站抓取的PlayStation奖杯指南数据，用于微调语言模型进行PlayStation奖杯指南问答任务。数据集包含原始抓取的指南数据（95K+条目）、处理后的训练数据（78个示例）以及从指南中提取的问答对（38个示例）。原始指南数据包括游戏标题、指南内容和元数据；训练数据采用指令/响应格式；问答对则是从指南中提取的具体问题和答案。数据集总共有5,287个游戏的指南，原始数据大小约为197 MB，平均指南长度约为5,000个标记。数据来源于PSNProfiles.com的奖杯指南部分，使用MIT许可证，适用于教育/研究目的。

This dataset contains trophy guide data scraped from PSNProfiles.com for fine-tuning language models on PlayStation trophy guide Q&A. The dataset includes raw scraped guide data (95K+ entries), processed training data (78 examples), and question-answer pairs extracted from guides (38 examples). Raw guide data includes game titles, guide content, and metadata; training data follows an instruction/response format; and QA pairs are specific questions and answers extracted from the guides. The dataset contains guides for 5,287 games, with a raw data size of ~197 MB and an average guide length of ~5,000 tokens. The data is sourced from the trophy guides section of PSNProfiles.com, licensed under MIT, and intended for educational/research purposes.