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 2502.17282cs.CLcs.AIcs.LG
收藏

能力指令调优:动态大语言模型路由的新范式

Capability Instruction Tuning: A New Paradigm for Dynamic LLM Routing

School of Artificial Intelligence, Nanjing University(南京大学人工智能学院);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)

Yi-Kai Zhang, De-Chuan Zhan, Han-Jia Ye

AI总结:

提出能力指令调优新范式,通过构建能力指令和模型能力编码器,实现无需候选推理的动态LLM路由,达到最先进性能。

摘要:

大型语言模型(LLMs)已展现出类人的指令遵循能力,尤其是那些超过1000亿参数的模型。一些较小的、资源友好的LLM的联合能力可以处理大多数大型LLM擅长的指令。在这项工作中,我们探讨如何为每条指令路由性能最佳的LLM,以实现更好的整体性能。我们开发了一种新范式,通过构建包含模型能力表示、用户指令和性能查询提示的能力指令来评估性能。为了从能力指令中学习,我们引入了一个新的端到端框架,称为基于能力测试的模型选择(Model-SAT),该框架根据不同模型擅长或难以处理的内容生成正负样本。Model-SAT使用一个模型能力编码器,将其模型表示扩展到轻量级LLM。我们的实验表明,Model-SAT能够理解候选模型的性能维度,并提供它们处理各种指令的能力概率。此外,在部署期间,新模型可以快速推断其在50个任务(每个任务20个样本)中的能力测试结果。Model-SAT在没有候选推理的情况下实现了最先进的模型路由,并在真实世界新模型发布场景中表现出色。代码可在https://github.com/Now-Join-Us/CIT-LLM-Routing获取。

英文摘要:

Large Language Models (LLMs) have demonstrated human-like instruction-following abilities, particularly those exceeding 100 billion parameters. The combined capability of some smaller, resource-friendly LLMs can address most of the instructions that larger LLMs excel at. In this work, we explore how to route the best-performing LLM for each instruction to achieve better overall performance. We develop a new paradigm, constructing capability instructions with model capability representation, user instruction, and performance inquiry prompts to assess the performance. To learn from capability instructions, we introduce a new end-to-end framework called Model Selection with Aptitude Test (Model-SAT), which generates positive and negative samples based on what different models perform well or struggle with. Model-SAT uses a model capability encoder that extends its model representation to a lightweight LLM. Our experiments show that Model-SAT understands the performance dimensions of candidate models and provides the probabilities of their capability to handle various instructions. Additionally, during deployment, a new model can quickly infer its aptitude test results across 50 tasks, each with 20 shots. Model-SAT performs state-of-the-art model routing without candidate inference and in real-world new model-released scenarios. The code is available at https://github.com/Now-Join-Us/CIT-LLM-Routing

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。

↑