OVM-process

The training dataset of GSM8K for process reward models in the paper [OVM, Outcome-supervised Value Models for Planning in Mathematical Reasoning](https://arxiv.org/pdf/2311.09724.pdf), where the responses were generated by llama2-7b and the labels were annotated by GPT-4.

Steps are split by the newlines in the response. `step_labels` indicates the logical correctness of steps, defined as "logically correct and it's based on accurate premises, not necessarily helps to solve the problem"; `step_labels_progress` indicates helpfulness of steps, defined as "logically correct, based on accurate premises, and helps to solve the problem".

本数据集为刊载于论文《OVM：数学推理规划中的结果监督价值模型》（https://arxiv.org/pdf/2311.09724.pdf）的面向过程奖励模型的GSM8K训练数据集，其模型生成的回复由llama2-7b产出，数据集标签由GPT-4标注。

数据集中的推理步骤以回复内的换行符进行分割。`step_labels`用于表征步骤的逻辑正确性，其定义为："逻辑正确且基于准确前提，未必有助于问题求解"；`step_labels_progress`用于表征步骤的解题辅助性，其定义为："逻辑正确且基于准确前提，同时有助于问题求解"。

搜集汇总

背景与挑战

背景概述

该数据集是论文'Outcome-supervised Value Models for Planning in Mathematical Reasoning'中用于过程奖励模型的GSM8K训练数据，由llama2-7b生成响应并由GPT-4标注标签。它包含步骤标签，用于评估步骤的逻辑正确性和对解决问题的帮助性。