arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2310.12426cs.CLcs.AI
收藏

MAF:用于改进大型语言模型推理的多方面反馈

MAF: Multi-Aspect Feedback for Improving Reasoning in Large Language Models

University of California, Santa Barbara(加州大学圣巴巴拉分校)

Deepak Nathani, David Wang, Liangming Pan, William Yang Wang

AI总结:

本文提出 Multi-Aspect Feedback 迭代精炼框架,通过集成多个专注特定错误类别的反馈模块(含冻结LM和外部工具),有效解决LM推理链中的多种错误,在数学推理和逻辑蕴含任务上分别实现高达20%和18%的相对提升。

摘要:

语言模型(LMs)在各种自然语言任务中表现出色。然而,在自然语言推理方面,LMs 仍面临挑战,如幻觉、生成错误的中间推理步骤以及出现数学错误。近期研究侧重于通过使用反馈进行自我改进来增强 LMs。尽管如此,依赖单一通用反馈源的现有方法无法解决 LM 生成的推理链中存在的多种错误类型。在这项工作中,我们提出了 Multi-Aspect Feedback,这是一个迭代精炼框架,它集成了多个反馈模块,包括冻结的 LMs 和外部工具,每个模块专注于特定的错误类别。我们的实验结果证明了我们的方法在解决 LM 生成的推理链中的若干错误方面的有效性,从而提高了 LM 在多项推理任务中的整体表现。我们观察到,在 Mathematical Reasoning 中相对改进高达 20%,在 Logical Entailment 中相对改进高达 18%。

英文摘要:

Language Models (LMs) have shown impressive performance in various natural language tasks. However, when it comes to natural language reasoning, LMs still face challenges such as hallucination, generating incorrect intermediate reasoning steps, and making mathematical errors. Recent research has focused on enhancing LMs through self-improvement using feedback. Nevertheless, existing approaches relying on a single generic feedback source fail to address the diverse error types found in LM-generated reasoning chains. In this work, we propose Multi-Aspect Feedback, an iterative refinement framework that integrates multiple feedback modules, including frozen LMs and external tools, each focusing on a specific error category. Our experimental results demonstrate the efficacy of our approach to addressing several errors in the LM-generated reasoning chain and thus improving the overall performance of an LM in several reasoning tasks. We see a relative improvement of up to 20% in Mathematical Reasoning and up to 18% in Logical Entailment.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑