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 2403.14541cs.CL
收藏

EDT:通过基于熵的动态温度采样改进大语言模型的生成

EDT: Improving Large Language Models' Generation by Entropy-based Dynamic Temperature Sampling

National Key Laboratory for Novel Software Technology, Nanjing University(南京大学计算机软件新技术国家重点实验室)

Shimao Zhang, Yu Bao, Shujian Huang

AI总结:

针对固定温度难以兼顾生成质量与多样性的问题,本文提出基于熵动态调整温度的EDT采样方法,并在4个生成基准上验证其显著优于现有策略。

摘要:

近来,大语言模型(LLM)在广泛的下游语言任务中展现出卓越性能。温度采样是LLM生成过程中常用的解码策略。然而,多数情况下使用的是固定温度参数,这在平衡生成质量与多样性时未必总是最佳选择。本文提出一种有效的基于熵的动态温度(Entropy-based Dynamic Temperature,EDT)采样方法,通过动态选择温度参数,在生成质量和多样性两方面实现更均衡的性能。此外,我们还展示了模型在4个不同生成基准上的性能,并进行了全面分析。实验表明,EDT在不同任务上均显著优于现有策略。

英文摘要:

Recently, Large Language Models (LLMs) have demonstrated outstanding performance across a wide range of downstream language tasks. Temperature sampling is a commonly used decoding strategy for LLMs' generation process. However, a fixed temperature parameter is used in most cases, which may not always be an optimal choice for balancing generation quality and diversity. In this paper, we propose an effective Entropy-based Dynamic Temperature (EDT) Sampling method, to achieve a more balanced performance in terms of both generation quality and diversity by dynamically selecting the temperature parameter. Additionally, we also show model performance and comprehensive analyses for 4 different generation benchmarks. Our experiments show that EDT significantly outperforms the existing strategies across different 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。

↑