arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.15526cs.AIcs.LG
收藏

特异性链:一种从大型语言模型中迭代提炼知识的细化方法

Chain-of-Specificity: An Iteratively Refining Method for Eliciting Knowledge from Large Language Models

Kuaishou Inc.(快手公司)

Kaiwen Wei, Jingyuan Zhang, Hongzhi Zhang, Fuzheng Zhang, Di Zhang, Li Jin, Yue Yu

AI总结:

本文提出Chain-of-Specificity方法,通过迭代强调输入指令中的特定约束并释放LLM内部知识来细化响应,实验证明其在生成内容特异性上优于现有方法,且蒸馏其响应可提升小模型遵循约束指令的能力。

摘要:

大型语言模型(LLMs)展现出卓越的生成能力,能够生成有价值的信息。尽管取得了这些进展,先前的研究发现LLMs有时难以遵守特定约束(例如,在特定地点或特定时间),有时甚至忽略这些约束,从而导致响应要么过于笼统,要么不能完全令人满意。现有方法试图通过分解或重写输入指令来解决这一问题,但它们在充分强调特定约束和释放潜在知识(例如,在软件开发情境下的编程知识)方面仍存在不足。针对这一问题,本文提出了一种简单而有效的方法,名为Chain-of-Specificity(CoS)。具体而言,CoS迭代地强调输入指令中的特定约束,释放LLMs内部的知识,并细化响应。在公开可用和自行构建的复杂数据集上进行的实验表明,CoS在增强生成内容方面优于现有方法,尤其是在特异性方面。此外,随着特定约束数量的增加,其他基线方法表现不佳,而CoS仍然表现良好。我们还表明,蒸馏由CoS生成的响应能够有效增强较小模型遵循约束指令的能力。本文的资源将被发布以供进一步研究。

英文摘要:

Large Language Models (LLMs) exhibit remarkable generative capabilities, enabling the generation of valuable information. Despite these advancements, previous research found that LLMs sometimes struggle with adhering to specific constraints (e.g., in specific place or at specific time), at times even overlooking them, which leads to responses that are either too generic or not fully satisfactory. Existing approaches attempted to address this issue by decomposing or rewriting input instructions, yet they fall short in adequately emphasizing specific constraints and in unlocking the underlying knowledge (e.g., programming within the context of software development). In response, this paper proposes a simple yet effective method named Chain-of-Specificity (CoS). Specifically, CoS iteratively emphasizes the specific constraints in the input instructions, unlocks knowledge within LLMs, and refines responses. Experiments conducted on publicly available and self-build complex datasets demonstrate that CoS outperforms existing methods in enhancing generated content especially for the specificity. Besides, as the number of specific constraints increase, other baselines falter, while CoS still performs well. Moreover, we show that distilling responses generated by CoS effectively enhances the ability of smaller models to follow the constrained instructions. Resources of this paper will be released for further research.

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。

↑