Avelina/UltraSteer-v0-flat
收藏Hugging Face2024-10-04 更新2024-12-14 收录
下载链接：
https://hf-mirror.com/datasets/Avelina/UltraSteer-v0-flat
下载链接
链接失效反馈官方服务：
资源简介：
UltraSteer是一个大规模的对话数据集，包含单轮和多轮对话，并带有细粒度的标签。这些标签由Nvidia的Llama2-13B-SteerLM-RM奖励模型生成，使用了NeMo Aligner框架。每个助手的回答都会被评分，评分标准包括质量、毒性、幽默、创造力、帮助性、正确性、连贯性、复杂性和冗长性。数据集来源于五个高质量的对齐数据集和一个聊天数据集，并进行了去重和过滤处理。
UltraSteer is a large-scale dataset of single- and multi-turn dialogue with fine-grained labels produced by Nvidias Llama2-13B-SteerLM-RM reward model using the NeMo Aligner framework. Each assistant turn is rated with attributes such as quality, toxicity, humor, creativity, helpfulness, correctness, coherence, complexity, and verbosity. The dataset is sourced from five high-quality alignment datasets and one chat dataset, with deduplication and filtering applied.
提供机构：
Avelina
