Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Lin, Yen-Ting, Jin, Di, Xu, Tengyu, Wu, Tianhao, Sukhbaatar, Sainbayar, Zhu, Chen, He, Yun, Chen, Yun-Nung, Weston, Jason, Tian, Yuandong, Rahnama, Arash, Wang, Sinong, Ma, Hao, Fang, Han
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2501.10799
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910789474975744
author Lin, Yen-Ting
Jin, Di
Xu, Tengyu
Wu, Tianhao
Sukhbaatar, Sainbayar
Zhu, Chen
He, Yun
Chen, Yun-Nung
Weston, Jason
Tian, Yuandong
Rahnama, Arash
Wang, Sinong
Ma, Hao
Fang, Han
author_facet Lin, Yen-Ting
Jin, Di
Xu, Tengyu
Wu, Tianhao
Sukhbaatar, Sainbayar
Zhu, Chen
He, Yun
Chen, Yun-Nung
Weston, Jason
Tian, Yuandong
Rahnama, Arash
Wang, Sinong
Ma, Hao
Fang, Han
contents Large language models (LLMs) have recently demonstrated remarkable success in mathematical reasoning. Despite progress in methods like chain-of-thought prompting and self-consistency sampling, these advances often focus on final correctness without ensuring that the underlying reasoning process is coherent and reliable. This paper introduces Step-KTO, a training framework that combines process-level and outcome-level binary feedback to guide LLMs toward more trustworthy reasoning trajectories. By providing binary evaluations for both the intermediate reasoning steps and the final answer, Step-KTO encourages the model to adhere to logical progressions rather than relying on superficial shortcuts. Our experiments on challenging mathematical benchmarks show that Step-KTO significantly improves both final answer accuracy and the quality of intermediate reasoning steps. For example, on the MATH-500 dataset, Step-KTO achieves a notable improvement in Pass@1 accuracy over strong baselines. These results highlight the promise of integrating stepwise process feedback into LLM training, paving the way toward more interpretable and dependable reasoning capabilities.
format Preprint
id arxiv_https___arxiv_org_abs_2501_10799
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Step-KTO: Optimizing Mathematical Reasoning through Stepwise Binary Feedback
Lin, Yen-Ting
Jin, Di
Xu, Tengyu
Wu, Tianhao
Sukhbaatar, Sainbayar
Zhu, Chen
He, Yun
Chen, Yun-Nung
Weston, Jason
Tian, Yuandong
Rahnama, Arash
Wang, Sinong
Ma, Hao
Fang, Han
Machine Learning
Artificial Intelligence
Large language models (LLMs) have recently demonstrated remarkable success in mathematical reasoning. Despite progress in methods like chain-of-thought prompting and self-consistency sampling, these advances often focus on final correctness without ensuring that the underlying reasoning process is coherent and reliable. This paper introduces Step-KTO, a training framework that combines process-level and outcome-level binary feedback to guide LLMs toward more trustworthy reasoning trajectories. By providing binary evaluations for both the intermediate reasoning steps and the final answer, Step-KTO encourages the model to adhere to logical progressions rather than relying on superficial shortcuts. Our experiments on challenging mathematical benchmarks show that Step-KTO significantly improves both final answer accuracy and the quality of intermediate reasoning steps. For example, on the MATH-500 dataset, Step-KTO achieves a notable improvement in Pass@1 accuracy over strong baselines. These results highlight the promise of integrating stepwise process feedback into LLM training, paving the way toward more interpretable and dependable reasoning capabilities.
title Step-KTO: Optimizing Mathematical Reasoning through Stepwise Binary Feedback
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2501.10799