argilla/DIBT_prompts_ranked_synthetic_n3_s1000

---
license: apache-2.0
dataset_info:
features:
- name: input
dtype: string
- name: quality
list:
- name: status
dtype: string
- name: user_id
dtype: string
- name: value
dtype: string
- name: metadata
dtype: string
- name: avg_rating
dtype: float64
- name: num_responses
dtype: int64
- name: agreement_ratio
dtype: float64
- name: raw_responses
sequence: int64
- name: kind
dtype: string
- name: cluster_description
dtype: string
- name: topic
dtype: string
- name: generation_model
sequence: string
- name: generation_prompt
list:
list:
- name: content
dtype: string
- name: role
dtype: string
- name: raw_generation_responses
sequence: string
- name: rating
sequence: float64
- name: rationale
sequence: string
- name: generations
sequence: 'null'
- name: hash
dtype: int64
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 124631137
num_examples: 10328
download_size: 25887169
dataset_size: 124631137
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
---

原始信息汇总

数据集概述

许可证

• Apache 2.0

数据集信息

特征

• input: 类型为字符串
• quality: 包含以下子特征
  • status: 类型为字符串
  • user_id: 类型为字符串
  • value: 类型为字符串
• metadata: 类型为字符串
• avg_rating: 类型为浮点数 (float64)
• num_responses: 类型为整数 (int64)
• agreement_ratio: 类型为浮点数 (float64)
• raw_responses: 类型为整数序列 (sequence: int64)
• kind: 类型为字符串
• cluster_description: 类型为字符串
• topic: 类型为字符串
• generation_model: 类型为字符串序列 (sequence: string)
• generation_prompt: 包含以下子特征
  • content: 类型为字符串
  • role: 类型为字符串
• raw_generation_responses: 类型为字符串序列 (sequence: string)
• rating: 类型为浮点数序列 (sequence: float64)
• rationale: 类型为字符串序列 (sequence: string)
• generations: 类型为空序列 (sequence: null)
• hash: 类型为整数 (int64)
• index_level_0: 类型为整数 (int64)

数据分割

• train: 包含10328个样本，占用124631137字节

数据大小

• 下载大小: 25887169字节
• 数据集大小: 124631137字节

配置

• default: 包含训练数据文件，路径为 data/train-*