andrewsiah/personalization_prompt_response_gemma_2b

---
dataset_info:
features:
- name: reward_1
dtype: float64
- name: reward_2
dtype: float64
- name: reward_3
dtype: float64
- name: reward_4
dtype: float64
- name: reward_5
dtype: float64
- name: reward_6
dtype: float64
- name: reward_7
dtype: float64
- name: reward_8
dtype: float64
- name: prompt
dtype: string
- name: subset
dtype: string
- name: id
dtype: int64
- name: response_1
dtype: string
- name: response_1_model
dtype: string
- name: response_2
dtype: string
- name: response_2_model
dtype: string
- name: response_3
dtype: string
- name: response_3_model
dtype: string
- name: response_4
dtype: string
- name: response_4_model
dtype: string
- name: response_5
dtype: string
- name: response_5_model
dtype: string
- name: response_6
dtype: string
- name: response_6_model
dtype: string
- name: response_7
dtype: string
- name: response_7_model
dtype: string
- name: response_8
dtype: string
- name: response_8_model
dtype: string
- name: rformatted_promptresponse_1
dtype: string
- name: rformatted_promptresponse_2
dtype: string
- name: rformatted_promptresponse_3
dtype: string
- name: rformatted_promptresponse_4
dtype: string
- name: rformatted_promptresponse_5
dtype: string
- name: rformatted_promptresponse_6
dtype: string
- name: rformatted_promptresponse_7
dtype: string
- name: rformatted_promptresponse_8
dtype: string
splits:
- name: train
num_bytes: 318180521
num_examples: 10431
download_size: 186360236
dataset_size: 318180521
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
---

This dataset is primarily used for evaluating and training models on reward mechanisms across multiple responses. It includes multiple reward features (reward_1 to reward_8), a prompt, subset, sample ID (id), and multiple responses (response_1 to response_8) along with their corresponding model information. Additionally, it contains formatted prompt-response pairs (rformatted_promptresponse_1 to rformatted_promptresponse_8). The dataset consists of a single training set (train) with 10431 samples, totaling 318180521 bytes in size.

原始信息汇总

数据集概述

数据集特征

• reward_1 至 reward_8: 数据类型为 float64。
• prompt: 数据类型为 string。
• subset: 数据类型为 string。
• id: 数据类型为 int64。
• response_1 至 response_8: 数据类型为 string。
• response_1_model 至 response_8_model: 数据类型为 string。
• rformatted_promptresponse_1 至 rformatted_promptresponse_8: 数据类型为 string。

数据集划分

• train: 包含 10431 个示例，数据大小为 318180521 字节。

数据集大小

• 下载大小: 186360236 字节。
• 数据集大小: 318180521 字节。