five

rishik12/gair_flip50

收藏
Hugging Face2024-03-13 更新2024-06-11 收录
下载链接:
https://hf-mirror.com/datasets/rishik12/gair_flip50
下载链接
链接失效反馈
官方服务:
资源简介:
--- configs: - config_name: default data_files: - split: train path: data/train-* - split: test path: data/test-* dataset_info: features: - name: query dtype: string - name: scenario_auto-j dtype: string - name: scenario_group dtype: string - name: response_1 struct: - name: content dtype: string - name: model dtype: string - name: num_words dtype: int64 - name: response_2 struct: - name: content dtype: string - name: model dtype: string - name: num_words dtype: int64 - name: gpt-4-turbo_reference dtype: string - name: clear intent dtype: string - name: explicitly express feelings dtype: string - name: explicit constraints sequence: string - name: explicit subjective stances sequence: string - name: explicit mistakes or biases sequence: string - name: preference_labels struct: - name: gpt-3.5-turbo-1106 dtype: string - name: gpt-4-1106-preview dtype: string - name: human dtype: string - name: llama-2-13b dtype: string - name: llama-2-13b-chat dtype: string - name: llama-2-70b dtype: string - name: llama-2-70b-chat dtype: string - name: llama-2-7b dtype: string - name: llama-2-7b-chat dtype: string - name: mistral-7b dtype: string - name: mistral-7b-instruct-v0.1 dtype: string - name: mistral-7b-instruct-v0.2 dtype: string - name: mistral-8x7b dtype: string - name: mistral-8x7b-instruct-v0.1 dtype: string - name: qwen-14b dtype: string - name: qwen-14b-chat dtype: string - name: qwen-72b dtype: string - name: qwen-72b-chat dtype: string - name: qwen-7b dtype: string - name: qwen-7b-chat dtype: string - name: tulu-2-dpo-13b dtype: string - name: tulu-2-dpo-70b dtype: string - name: tulu-2-dpo-7b dtype: string - name: vicuna-13b-v1.5 dtype: string - name: vicuna-7b-v1.5 dtype: string - name: wizardLM-13b-v1.2 dtype: string - name: wizardLM-70b-v1.0 dtype: string - name: yi-34b dtype: string - name: yi-34b-chat dtype: string - name: yi-6b dtype: string - name: yi-6b-chat dtype: string - name: zephyr-7b-alpha dtype: string - name: zephyr-7b-beta dtype: string - name: basic_response_1 struct: - name: admit limitations or mistakes dtype: int64 - name: authoritative tone dtype: int64 - name: clear and understandable dtype: int64 - name: complex word usage and sentence structure dtype: int64 - name: friendly dtype: int64 - name: funny and humorous dtype: int64 - name: grammar, spelling, punctuation, and code-switching dtype: int64 - name: harmlessness dtype: int64 - name: information richness without considering inaccuracy dtype: int64 - name: innovative and novel dtype: int64 - name: interactive dtype: int64 - name: metaphors, personification, similes, hyperboles, irony, parallelism dtype: int64 - name: persuade user dtype: int64 - name: polite dtype: int64 - name: relevance without considering inaccuracy dtype: int64 - name: repetitive dtype: int64 - name: step by step solution dtype: int64 - name: use of direct and explicit supporting materials dtype: int64 - name: use of informal expressions dtype: int64 - name: well formatted dtype: int64 - name: basic_response_2 struct: - name: admit limitations or mistakes dtype: int64 - name: authoritative tone dtype: int64 - name: clear and understandable dtype: int64 - name: complex word usage and sentence structure dtype: int64 - name: friendly dtype: int64 - name: funny and humorous dtype: int64 - name: grammar, spelling, punctuation, and code-switching dtype: int64 - name: harmlessness dtype: int64 - name: information richness without considering inaccuracy dtype: int64 - name: innovative and novel dtype: int64 - name: interactive dtype: int64 - name: metaphors, personification, similes, hyperboles, irony, parallelism dtype: int64 - name: persuade user dtype: int64 - name: polite dtype: int64 - name: relevance without considering inaccuracy dtype: int64 - name: repetitive dtype: int64 - name: step by step solution dtype: int64 - name: use of direct and explicit supporting materials dtype: int64 - name: use of informal expressions dtype: int64 - name: well formatted dtype: int64 - name: errors_response_1 struct: - name: applicable or not dtype: string - name: errors list: - name: brief description dtype: string - name: severity dtype: string - name: type dtype: string - name: errors_response_2 struct: - name: applicable or not dtype: string - name: errors list: - name: brief description dtype: string - name: severity dtype: string - name: type dtype: string - name: query-specific_response_1 struct: - name: clarify user intent dtype: float64 - name: correcting explicit mistakes or biases sequence: string - name: satisfying explicit constraints sequence: string - name: showing empathetic dtype: float64 - name: supporting explicit subjective stances sequence: string - name: query-specific_response_2 struct: - name: clarify user intent dtype: float64 - name: correcting explicit mistakes or biases sequence: string - name: satisfying explicit constraints sequence: string - name: showing empathetic dtype: float64 - name: supporting explicit subjective stances sequence: string - name: noise dtype: int64 splits: - name: train num_bytes: 27659291 num_examples: 5240 - name: test num_bytes: 27659291 num_examples: 5240 download_size: 26253307 dataset_size: 55318582 --- # Dataset Card for "gair_flip50" [More Information needed](https://github.com/huggingface/datasets/blob/main/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
提供机构:
rishik12
原始信息汇总

数据集概述

数据集配置

  • 默认配置
    • 训练集:路径为 data/train-*
    • 测试集:路径为 data/test-*

数据集信息

特征

  • query:字符串类型
  • scenario_auto-j:字符串类型
  • scenario_group:字符串类型
  • response_1:结构体,包含以下字段:
    • content:字符串类型
    • model:字符串类型
    • num_words:64位整数类型
  • response_2:结构体,包含以下字段:
    • content:字符串类型
    • model:字符串类型
    • num_words:64位整数类型
  • gpt-4-turbo_reference:字符串类型
  • clear intent:字符串类型
  • explicitly express feelings:字符串类型
  • explicit constraints:字符串序列类型
  • explicit subjective stances:字符串序列类型
  • explicit mistakes or biases:字符串序列类型
  • preference_labels:结构体,包含以下字段:
    • gpt-3.5-turbo-1106:字符串类型
    • gpt-4-1106-preview:字符串类型
    • human:字符串类型
    • llama-2-13b:字符串类型
    • llama-2-13b-chat:字符串类型
    • llama-2-70b:字符串类型
    • llama-2-70b-chat:字符串类型
    • llama-2-7b:字符串类型
    • llama-2-7b-chat:字符串类型
    • mistral-7b:字符串类型
    • mistral-7b-instruct-v0.1:字符串类型
    • mistral-7b-instruct-v0.2:字符串类型
    • mistral-8x7b:字符串类型
    • mistral-8x7b-instruct-v0.1:字符串类型
    • qwen-14b:字符串类型
    • qwen-14b-chat:字符串类型
    • qwen-72b:字符串类型
    • qwen-72b-chat:字符串类型
    • qwen-7b:字符串类型
    • qwen-7b-chat:字符串类型
    • tulu-2-dpo-13b:字符串类型
    • tulu-2-dpo-70b:字符串类型
    • tulu-2-dpo-7b:字符串类型
    • vicuna-13b-v1.5:字符串类型
    • vicuna-7b-v1.5:字符串类型
    • wizardLM-13b-v1.2:字符串类型
    • wizardLM-70b-v1.0:字符串类型
    • yi-34b:字符串类型
    • yi-34b-chat:字符串类型
    • yi-6b:字符串类型
    • yi-6b-chat:字符串类型
    • zephyr-7b-alpha:字符串类型
    • zephyr-7b-beta:字符串类型
  • basic_response_1:结构体,包含以下字段:
    • admit limitations or mistakes:64位整数类型
    • authoritative tone:64位整数类型
    • clear and understandable:64位整数类型
    • complex word usage and sentence structure:64位整数类型
    • friendly:64位整数类型
    • funny and humorous:64位整数类型
    • grammar, spelling, punctuation, and code-switching:64位整数类型
    • harmlessness:64位整数类型
    • information richness without considering inaccuracy:64位整数类型
    • innovative and novel:64位整数类型
    • interactive:64位整数类型
    • metaphors, personification, similes, hyperboles, irony, parallelism:64位整数类型
    • persuade user:64位整数类型
    • polite:64位整数类型
    • relevance without considering inaccuracy:64位整数类型
    • repetitive:64位整数类型
    • step by step solution:64位整数类型
    • use of direct and explicit supporting materials:64位整数类型
    • use of informal expressions:64位整数类型
    • well formatted:64位整数类型
  • basic_response_2:结构体,包含以下字段:
    • admit limitations or mistakes:64位整数类型
    • authoritative tone:64位整数类型
    • clear and understandable:64位整数类型
    • complex word usage and sentence structure:64位整数类型
    • friendly:64位整数类型
    • funny and humorous:64位整数类型
    • grammar, spelling, punctuation, and code-switching:64位整数类型
    • harmlessness:64位整数类型
    • information richness without considering inaccuracy:64位整数类型
    • innovative and novel:64位整数类型
    • interactive:64位整数类型
    • metaphors, personification, similes, hyperboles, irony, parallelism:64位整数类型
    • persuade user:64位整数类型
    • polite:64位整数类型
    • relevance without considering inaccuracy:64位整数类型
    • repetitive:64位整数类型
    • step by step solution:64位整数类型
    • use of direct and explicit supporting materials:64位整数类型
    • use of informal expressions:64位整数类型
    • well formatted:64位整数类型
  • errors_response_1:结构体,包含以下字段:
    • applicable or not:字符串类型
    • errors:列表,包含以下字段:
      • brief description:字符串类型
      • severity:字符串类型
      • type:字符串类型
  • errors_response_2:结构体,包含以下字段:
    • applicable or not:字符串类型
    • errors:列表,包含以下字段:
      • brief description:字符串类型
      • severity:字符串类型
      • type:字符串类型
  • query-specific_response_1:结构体,包含以下字段:
    • clarify user intent:64位浮点数类型
    • correcting explicit mistakes or biases:字符串序列类型
    • satisfying explicit constraints:字符串序列类型
    • showing empathetic:64位浮点数类型
    • supporting explicit subjective stances:字符串序列类型
  • query-specific_response_2:结构体,包含以下字段:
    • clarify user intent:64位浮点数类型
    • correcting explicit mistakes or biases:字符串序列类型
    • satisfying explicit constraints:字符串序列类型
    • showing empathetic:64位浮点数类型
    • supporting explicit subjective stances:字符串序列类型
  • noise:64位整数类型

数据集分割

  • 训练集
    • 字节数:27659291
    • 样本数:5240
  • 测试集
    • 字节数:27659291
    • 样本数:5240

数据集大小

  • 下载大小:26253307 字节
  • 数据集大小:55318582 字节
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作