Optimizing Chain-of-Thought Reasoning: Tackling Arranging Bottleneck via Plan Augmentation

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Qiu, Yuli, Yao, Jiashu, Huang, Heyan, Guo, Yuhang
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913559282188288
author Qiu, Yuli
Yao, Jiashu
Huang, Heyan
Guo, Yuhang
author_facet Qiu, Yuli
Yao, Jiashu
Huang, Heyan
Guo, Yuhang
contents Multi-step reasoning ability of large language models is crucial in tasks such as math and tool utilization. Current researches predominantly focus on enhancing model performance in these multi-step reasoning tasks through fine-tuning with Chain-of-Thought (CoT) steps, yet these methods tend to be heuristic, without exploring nor resolving the bottleneck. In this study, we subdivide CoT reasoning into two parts: arranging and executing, and identify that the bottleneck of models mainly lies in arranging rather than executing. Based on this finding, we propose a plan-based training and reasoning method that guides models to generate arranging steps through abstract plans. We experiment on both math (GSM8k) and tool utilization (ToolBench) benchmarks. Results show that compared to fine-tuning directly with CoT data, our approach achieves a better performance on alleviating arranging bottleneck, particularly excelling in long-distance reasoning generalization.
format Preprint
id arxiv_https___arxiv_org_abs_2410_16812
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Optimizing Chain-of-Thought Reasoning: Tackling Arranging Bottleneck via Plan Augmentation
Qiu, Yuli
Yao, Jiashu
Huang, Heyan
Guo, Yuhang
Computation and Language
Multi-step reasoning ability of large language models is crucial in tasks such as math and tool utilization. Current researches predominantly focus on enhancing model performance in these multi-step reasoning tasks through fine-tuning with Chain-of-Thought (CoT) steps, yet these methods tend to be heuristic, without exploring nor resolving the bottleneck. In this study, we subdivide CoT reasoning into two parts: arranging and executing, and identify that the bottleneck of models mainly lies in arranging rather than executing. Based on this finding, we propose a plan-based training and reasoning method that guides models to generate arranging steps through abstract plans. We experiment on both math (GSM8k) and tool utilization (ToolBench) benchmarks. Results show that compared to fine-tuning directly with CoT data, our approach achieves a better performance on alleviating arranging bottleneck, particularly excelling in long-distance reasoning generalization.
title Optimizing Chain-of-Thought Reasoning: Tackling Arranging Bottleneck via Plan Augmentation
topic Computation and Language
url https://arxiv.org/abs/2410.16812