arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2405.20701cs.CLcs.AI
收藏

揭示LLMs的词汇敏感性:用于提示增强的组合优化

Unveiling the Lexical Sensitivity of LLMs: Combinatorial Optimization for Prompt Enhancement

Institute of Information Engineering, Chinese Academy of Sciences(中国科学院信息工程研究所);School of Cyber Security, University of Chinese Academy of Sciences(中国科学院大学网络空间安全学院)

Pengwei Zhan, Zhen Xu, Qian Tan, Jie Song, Ru Xie

AI总结:

本文揭示了LLMs对任务指令中微小词汇变化过度敏感的问题,并提出黑盒组合优化框架COPLE,通过代理任务反馈迭代优化提示词汇,有效恢复模型在指令遵循和下游任务中的能力。

摘要:

大型语言模型(LLMs)展现出卓越的指令遵循能力,能够完成各种下游任务。尽管这种令人印象深刻的能力使LLMs成为灵活的任务解决者,但它们在解决任务时的性能也严重依赖于指令。在本文中,我们揭示了LLMs对任务指令中的词汇变化过度敏感,即使这些变化对人类来说是难以察觉的。通过向模型提供邻域指令——这些指令在潜在表示空间中紧密相邻,且仅相差一个语义相似的词——下游任务的性能可能会大不相同。基于这一特性,我们提出了一种用于提示词汇增强的黑盒组合优化框架(COPLE)。COPLE根据一批代理任务的反馈,使用与单词影响相关的搜索策略进行迭代词汇优化。实验表明,即使是当前基准测试中广泛使用的人工提示也受到模型词汇敏感性的影响,而COPLE在指令遵循和解决下游任务两方面都恢复了下降的模型能力。

英文摘要:

Large language models (LLMs) demonstrate exceptional instruct-following ability to complete various downstream tasks. Although this impressive ability makes LLMs flexible task solvers, their performance in solving tasks also heavily relies on instructions. In this paper, we reveal that LLMs are over-sensitive to lexical variations in task instructions, even when the variations are imperceptible to humans. By providing models with neighborhood instructions, which are closely situated in the latent representation space and differ by only one semantically similar word, the performance on downstream tasks can be vastly different. Following this property, we propose a black-box Combinatorial Optimization framework for Prompt Lexical Enhancement (COPLE). COPLE performs iterative lexical optimization according to the feedback from a batch of proxy tasks, using a search strategy related to word influence. Experiments show that even widely-used human-crafted prompts for current benchmarks suffer from the lexical sensitivity of models, and COPLE recovers the declined model ability in both instruct-following and solving downstream tasks.

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。

↑