Guardado en:
Detalles Bibliográficos
Autores principales: Gao, Bofei, Cai, Zefan, Xu, Runxin, Wang, Peiyi, Zheng, Ce, Lin, Runji, Lu, Keming, Liu, Dayiheng, Zhou, Chang, Xiao, Wen, Hu, Junjie, Liu, Tianyu, Chang, Baobao
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2406.14024
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866914977494859776
author Gao, Bofei
Cai, Zefan
Xu, Runxin
Wang, Peiyi
Zheng, Ce
Lin, Runji
Lu, Keming
Liu, Dayiheng
Zhou, Chang
Xiao, Wen
Hu, Junjie
Liu, Tianyu
Chang, Baobao
author_facet Gao, Bofei
Cai, Zefan
Xu, Runxin
Wang, Peiyi
Zheng, Ce
Lin, Runji
Lu, Keming
Liu, Dayiheng
Zhou, Chang
Xiao, Wen
Hu, Junjie
Liu, Tianyu
Chang, Baobao
contents In recent progress, mathematical verifiers have achieved success in mathematical reasoning tasks by validating the correctness of solutions generated by policy models. However, existing verifiers are trained with binary classification labels, which are not informative enough for the model to accurately assess the solutions. To mitigate the aforementioned insufficiency of binary labels, we introduce step-wise natural language feedback as rationale labels, that is, the correctness of each step and the detailed explanations. In this paper, we propose Math-Minos, a natural language feedback-enhanced verifier by constructing automatically generated training data and a two-stage training paradigm for effective training and efficient inference. Our experiments reveal that a small set of natural language feedback can significantly boost the performance of the verifier in both verification and reinforcement learning. We have released the code and data for further exploration.
format Preprint
id arxiv_https___arxiv_org_abs_2406_14024
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle LLM Critics Help Catch Bugs in Mathematics: Towards a Better Mathematical Verifier with Natural Language Feedback
Gao, Bofei
Cai, Zefan
Xu, Runxin
Wang, Peiyi
Zheng, Ce
Lin, Runji
Lu, Keming
Liu, Dayiheng
Zhou, Chang
Xiao, Wen
Hu, Junjie
Liu, Tianyu
Chang, Baobao
Computation and Language
In recent progress, mathematical verifiers have achieved success in mathematical reasoning tasks by validating the correctness of solutions generated by policy models. However, existing verifiers are trained with binary classification labels, which are not informative enough for the model to accurately assess the solutions. To mitigate the aforementioned insufficiency of binary labels, we introduce step-wise natural language feedback as rationale labels, that is, the correctness of each step and the detailed explanations. In this paper, we propose Math-Minos, a natural language feedback-enhanced verifier by constructing automatically generated training data and a two-stage training paradigm for effective training and efficient inference. Our experiments reveal that a small set of natural language feedback can significantly boost the performance of the verifier in both verification and reinforcement learning. We have released the code and data for further exploration.
title LLM Critics Help Catch Bugs in Mathematics: Towards a Better Mathematical Verifier with Natural Language Feedback
topic Computation and Language
url https://arxiv.org/abs/2406.14024