YYYYYYibo/ultrafeedback_binarized_gshf_lora_train_part_2
收藏Hugging Face2024-07-13 更新2024-07-22 收录
下载链接:
https://hf-mirror.com/datasets/YYYYYYibo/ultrafeedback_binarized_gshf_lora_train_part_2
下载链接
链接失效反馈官方服务:
资源简介:
该数据集包含多个特征,如prompt、prompt_id、messages、score_chosen、score_rejected、reference_response、chosen和rejected。每个特征都有其特定的数据类型,例如string或float64。数据集分为多个部分,例如train_prefs,包含特定数量的字节和示例。此外,数据集有特定的配置,如default,指定了数据文件的路径。整体上,数据集似乎用于处理和评估对话系统中的反馈和选择。
The dataset includes multiple features such as prompt, prompt_id, messages, score_chosen, score_rejected, reference_response, chosen, and rejected. Each feature has its specific data type, such as string or float64. The dataset is divided into parts, such as train_prefs, containing specific numbers of bytes and examples. Additionally, the dataset has specific configurations, such as default, specifying the paths of data files. Overall, the dataset appears to be used for processing and evaluating feedback and choices in dialogue systems.
提供机构:
YYYYYYibo
原始信息汇总
数据集概述
数据集信息
特征
- prompt: 字符串类型
- prompt_id: 字符串类型
- messages: 列表类型
- content: 字符串类型
- role: 字符串类型
- score_chosen: 浮点数类型
- score_rejected: 浮点数类型
- reference_response: 字符串类型
- chosen: 列表类型
- content: 字符串类型
- role: 字符串类型
- rejected: 列表类型
- content: 字符串类型
- role: 字符串类型
数据分割
- train_prefs:
- num_bytes: 172683063
- num_examples: 20000
数据集大小
- download_size: 93405525
- dataset_size: 172683063
配置
- config_name: default
- data_files:
- split: train_prefs
- path: data/train_prefs-*
- data_files:



