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 2510.16686cs.CL
收藏

探究大语言模型推理依据对自然语言理解的影响

Investigating the Impact of Rationales for LLMs on Natural Language Understanding

Wenhang Shi, Shuqing Bian, Yiren Chen, Xinyi Zhang, Zhe Zhao, Pengfei Hu, Wei Lu, Xiaoyong Du

AI总结:

本文构建带推理依据的NLU数据集NLURC并评估多种增强方法,发现CoT收益随模型规模由负转正,特定训练方法可让小模型媲美十倍规模模型。

摘要:

思维链(Chain-of-thought,CoT)推理依据通过提供逐步推理来得出最终答案,在推理和训练两方面都能使大语言模型(LLM)受益。引入推理依据——无论是在推理阶段于回答之前生成推理依据,还是在训练阶段将其置于原始答案之前或之后——都能显著提升模型在数学、符号和常识推理任务上的表现。然而,大多数研究关注推理依据在这些推理任务中的作用,忽视了其对自然语言理解(NLU)任务等其他重要任务的潜在影响。在本研究中,我们提出问题:推理依据能否同样使NLU任务受益?为进行系统性探索,我们构建了NLURC,这是一个全面、高质量且带有推理依据的NLU数据集集合,并开发了多种推理依据增强方法。通过利用该数据集探索这些方法在NLU任务上的适用性,我们发现了若干可能令人惊讶的结果:(1)随着模型规模增大,CoT推理会从阻碍NLU表现转变为超越直接标签预测,表明二者呈正相关。(2)大多数推理依据增强训练方法的表现不如仅使用标签的训练,但有一种专门设计的方法能够持续取得提升。(3)使用推理依据训练的LLM在未见过的NLU任务上获得显著性能提升,可与规模十倍于自身的模型相媲美,同时提供与商业LLM相当的可解释性。

英文摘要:

Chain-of-thought (CoT) rationales, which provide step-by-step reasoning to derive final answers, benefit LLMs in both inference and training. Incorporating rationales, either by generating them before answering during inference, or by placing them before or after the original answers during training - significantly improves model performance on mathematical, symbolic and commonsense reasoning tasks. However, most work focuses on the role of rationales in these reasoning tasks, overlooking their potential impact on other important tasks like natural language understanding (NLU) tasks. In this work, we raise the question: Can rationales similarly benefit NLU tasks? To conduct a systematic exploration, we construct NLURC, a comprehensive and high-quality NLU dataset collection with rationales, and develop various rationale-augmented methods. Through exploring the applicability of these methods on NLU tasks using the dataset, we uncover several potentially surprising findings: (1) CoT inference shifts from hindering NLU performance to surpassing direct label prediction as model size grows, indicating a positive correlation. (2) Most rationale-augmented training methods perform worse than label-only training, with one specially designed method consistently achieving improvements. (3) LLMs trained with rationales achieve significant performance gains on unseen NLU tasks, rivaling models ten times their size, while delivering interpretability on par with commercial LLMs.

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。

↑