mlfoundations-dev/instruction_filtering_askllm_seed_data_math
收藏Hugging Face2025-02-14 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/mlfoundations-dev/instruction_filtering_askllm_seed_data_math
下载链接
链接失效反馈官方服务：
资源简介：
这是一个包含对话数据的训练集，数据集包含多个字段，如指令种子、来源、是否被使用、分类器推理、原始行索引、70B蒸馏模型响应和对话。对话字段又包括对话来源和对话内容。训练集大小为85,122,493字节，共有2,500个样本。
This is a training dataset containing conversational data, which includes multiple fields such as instruction seed, source, to be used, classifier reasoning, original row index, 70B distillation model response, and conversation. The conversation field includes conversation source and conversation content. The training set is 85,122,493 bytes in size and contains 2,500 samples.
提供机构：
mlfoundations-dev
